Custom Resource options reference¶

Percona Operator for MySQL uses Custom Resources to manage options for the various components of the cluster.

PerconaServerMySQL Custom Resource with options for the cluster,
PerconaServerMySQLClusterSet Custom Resource with options for InnoDB ClusterSet — see ClusterSet Resource options and cross-site replication,
PerconaServerMySQLBackup Custom Resource contains options for Percona XtraBackup used to backup Percona Server for MySQL
PerconaServerMySQLRestore Custom Resource contains options for restoring Percona Server for MySQL from backups.

For status fields and conditions, see Custom resource statuses.

PerconaServerMySQL Custom Resource options¶

Percona Server for MySQL managed by the Operator is configured via the spec section of the deploy/cr.yaml file.

The metadata part of PerconaServerMySQL Custom Resource contains the following keys:

name (ps-cluster1 by default) sets the name of your Percona Server for MySQL cluster; it should include only URL-compatible characters , not exceed 22 characters, start with an alphabetic character, and end with an alphanumeric character;
finalizers subsection:
- percona.com/delete-mysql-pods-in-order if present, activates the Finalizer which controls the proper Pods deletion order in case of the cluster deletion event (on by default).
- percona.com/delete-mysql-pvc if present, activates the Finalizer which deletes Persistent Volume Claims for Percona Server for MySQL Pods after the cluster deletion event (off by default). It also triggers deletion of user Secrets.
- percona.com/delete-ssl if present, activates the Finalizer which deletes objects, created for SSL (Secret, certificate, and issuer) after the cluster deletion event (off by default).

The top-level spec elements of the deploy/cr.yaml are the following ones:

Toplevel `spec` elements¶

The spec part of the deploy/cr.yaml file contains the following:

`crVersion`¶

Version of the Operator the Custom Resource belongs to.

Value type	Example
string	`1.3.0`

`metadata.annotations`¶

The Kubernetes annotations metadata that you can set at a global level for all resources created by the Operator.

Value type	Example
label	`example-annotation: value`

`metadata.labels`¶

The Kubernetes labels metadata that you can set at a global level for all resources created by the Operator.

Value type	Example
label	`example-label: value`

`pause`¶

Pause/resume: setting it to true gracefully stops the cluster, and setting it to false after shut down starts the cluster back.

Value type	Example
boolean	`false`

`enableVolumeExpansion`¶

Enables or disables storage resizing with volume expansion.

Value type	Example
boolean	`false`

This option is deprecated and will be removed in version 1.5.0. Use the storageScaling.enableVolumeScaling option instead.

Storage scaling section¶

The storageScaling section in the deploy/cr.yaml file controls volume expansion and automatic storage resizing for MySQL PVCs. See Scale storage for usage details.

`storageScaling.enableVolumeScaling`¶

Enables or disables storage scaling / volume expansion with Volume Expansion capability.

Value type	Example
boolean	`false`

`storageScaling.enableExternalAutoscaling`¶

Enables or disables the use of external volume autoscaler. When disabled, the Operator uses its own expansion logic with Volume Expansion capability.

Value type	Example
boolean	`false`

`storageScaling.autoscaling.enabled`¶

Enables automatic storage resizing based on user-defined thresholds. Read more about this feature in Automatic storage resizing.

Value type	Example
boolean	`false`

`storageScaling.autoscaling.triggerThresholdPercent`¶

The percentage of the storage usage that triggers automatic resizing. Valid values are from 50 to 95. Default value is 80.

Value type	Example
int	`80`

`storageScaling.autoscaling.growthStep`¶

The amount to increase the storage during automatic resizing. Default value is 2Gi.

Value type	Example
string	`2Gi`

`storageScaling.autoscaling.maxSize`¶

The maximum size to which storage can be automatically resized. Autoscaling stops when storage reaches this limit. Must be at least 1Gi.

Value type	Example
string	`10Gi`

`initContainer.Image`¶

An alternative init image for the Operator.

Value type	Example
string	`perconalab/percona-server-mysql-operator:1.3.0`

`initContainer.containerSecurityContext`¶

A custom Kubernetes Security Context for a Container used for the initial Operator installation.

Value type	Example
subdoc	privileged: false runAsUser: 1001 runAsGroup: 1001

`initContainer.resources.requests.memory`¶

The Kubernetes memory requests for an image used for the initial Operator installation.

Value type	Example
string	`100M`

`initContainer.resources.requests.cpu`¶

Kubernetes CPU requests for an image used for the initial Operator installation.

Value type	Example
string	`100m`

`initContainer.resources.limits.memory`¶

Kubernetes memory limits for an image used for the initial Operator installation.

Value type	Example
string	`200M`

`initContainer.resources.limits.cpu`¶

Kubernetes CPU limits for an image used for the initial Operator installation.

Value type	Example
string	`200m`

`secretsName`¶

A name for users secrets. When undefined, the Operator creates the Secrets object named in the format <cluster-name>-secrets. Otherwise, it uses the provided name.

Value type	Example
string	`ps-cluster1-secrets`

`clusterServiceDNSSuffix`¶

A custom cluster domain suffix the Operator uses when constructing internal service hostnames. Use this when the Operator runs in a vcluster or in a cluster with a custom DNS domain so it can correctly resolve services. See Configure DNS suffix for service discovery for details.

Value type	Example
string	`cluster.local`

`sslSecretName`¶

A secret with the TLS certificate used by the cluster, see Transport Layer Security (TLS) for details. When undefined, the Operator creates the Secrets object named <cluster-name>-ssl. Otherwise, it uses the provided name. If this Secret already exists and was created by you, the Operator uses it and does not request a certificate from cert-manager.

Value type	Example
string	`ps-cluster1-ssl`

`ignoreAnnotations`¶

The list of annotations to be ignored by the Operator.

Value type	Example
subdoc	`service.beta.kubernetes.io/aws-load-balancer-backend-protocol`

`ignoreLabels`¶

The list of labels to be ignored by the Operator.

Value type	Example
subdoc	`rack`

`updateStrategy`¶

A strategy the Operator uses for upgrades.

Value type	Example
string	`SmartUpdate`

`pause`¶

Pause/resume: setting it to true gracefully stops the cluster, and setting it to false after shut down starts the cluster back.

Value type	Example
boolean	`false`

`allowUnsafeConfigurations`¶

Prevents users from configuring a cluster with unsafe parameters such as starting a group replication cluster with less than 3 or more than 9 Percona Server for MySQL instances. This option is deprecated and will be removed in future releases. Use unsafeFlags subsection instead. Setting allowUnsafeConfigurations won’t have any effect with the Operator version 0.8.0 and newer, and upgrading existing clusters with allowUnsafeConfigurations=true will cause everything under the unsafeFlags subsection set to true.

Value type	Example
boolean	`false`

Unsafe flags section¶

The unsafeFlags section in the deploy/cr.yaml file contains various configuration options to prevent users from configuring a cluster with unsafe parameters. *

Warning

Once unsafe configurations are enabled, you cannot revert these settings simply by setting the same keys back to false. Any such changes will be ignored, so enable these options with caution.

`unsafeFlags.asyncFailoverWithPossibleDataLoss`¶

Allows users to disable backups in a cluster with asynchronous replication. Without this flag, the Operator rejects such a Custom Resource, because a zero-data-loss failover reads the failed primary’s binary logs from the xtrabackup sidecar, and the sidecar runs only when backups are enabled. With backups disabled, a failover promotes a replica without recovering the transactions that the failed primary has not yet replicated. The Operator records the FailoverUnprotected event.

Value type	Example
boolean	`false`

`unsafeFlags.backupNonReadyCluster`¶

Allows the Operator to run a backup from a healthy MySQL Pod when the cluster is not yet ready.

Value type	Example
boolean	`false`

`unsafeFlags.mysqlSize`¶

Allows a cluster size that is outside the safe defaults for the replication type:

Asynchronous (async): a single MySQL instance (mysql.size: 1). Size 2 or greater is allowed without this flag starting with Operator 1.3.0, including even counts.
Group Replication (group-replication): an even instance count, or more than 9 instances.

Value type	Example
boolean	`false`

`unsafeFlags.proxy`¶

Allows users to configure a cluster with disabled proxy (both HAProxy and Router).

Value type	Example
boolean	`false`

`unsafeFlags.proxySize`¶

Allows users to set proxy (HAProxy or Router) size to a value less than 2 Pods.

Value type	Example
boolean	`false`

`unsafeFlags.orchestrator`¶

Allows users to configure a cluster with disabled Orchestrator even if asynchronous replication is turned on.

Value type	Example
boolean	`false`

`unsafeFlags.orchestratorSize`¶

Allows users to set orchestrator.size option to a value less than the minimum safe size (3).

Value type	Example
boolean	`false`

Extended cert-manager configuration section¶

The tls section in the deploy/cr.yaml file contains options that control how the Operator requests TLS certificates from cert-manager.

`tls.SANs`¶

Additional DNS names (subject alternative names) to add to the TLS certificate that cert-manager issues.

Value type	Example
subdoc	`mysql-1.example.com`

`tls.issuerConf.name`¶

The name of the cert-manager issuer that signs the cluster TLS certificate.

When set, the Operator creates a Certificate that references this issuer and does not create its own CA chain. Use this to integrate with an existing organizational PKI. See Use an existing ClusterIssuer and Use an existing namespace-scoped Issuer.

Value type	Example
string	`special-selfsigned-issuer`

`tls.issuerConf.kind`¶

The cert-manager issuer type referenced by the cluster Certificate.

Supported values:

Issuer (default) — namespace-scoped issuer in the database namespace.
ClusterIssuer — cluster-scoped issuer. Use this when your platform team manages a cluster-wide issuer. See Use an existing ClusterIssuer.

Value type	Example
string	`ClusterIssuer`

`tls.issuerConf.group`¶

The cert-manager API group for the issuer referenced in issuerConf. Use cert-manager.io for built-in cert-manager issuers.

Value type	Example
string	`cert-manager.io`

Upgrade options section¶

The upgradeOptions section in the deploy/cr.yaml file contains various configuration options to control Percona Server for MySQL version choice at the deployment time and during upgrades.

`upgradeOptions.versionServiceEndpoint`¶

The Version Service URL used to check versions compatibility for upgrade.

Value type	Example
string	`https://check.percona.com`

`upgradeOptions.apply`¶

Specifies how images are picked up from the version service on initial start by the Operator. Never or Disabled will completely disable querying version service for images, otherwise it can be set to Latest or Recommended or to a specific version string of Percona Server for MySQL (e.g. 8.0.32-24) that is wished to be version-locked (so that the user can control the version running) |

Value type	Example
string	`Disabled`

Percona Server for MySQL section¶

The mysql section in the deploy/cr.yaml file contains general configuration options for the Percona Server for MySQL.

`mysql.clusterType`¶

The cluster type: async for Asynchronous replication , group-replication for Group Replication .

Starting with Operator version 1.3.0, you can change this value on a running cluster. The Operator converts the topology in place. The switch causes downtime. See Change replication type to learn more.

Value type	Example
int	`group-replication`

`mysql.bootstrap.mode`¶

Controls whether the Operator forms a Group Replication cluster automatically on first start. Use manual for replica sites in cross-site replication.

Value type	Example
string	`manual`

Supported values:

Value	Meaning
`auto` (default)	Pod-0 runs the `dba.createCluster()` function when no peers exist. The same behavior is when the value for `bootstrap` is not set.
`manual`	Pod-0 configures MySQL for Group Replication but does not create a cluster. Pod-0 stays in the `NotReady` state until the ClusterSet controller runs `createReplicaCluster()`. Pod-1 and above do not start until Pod-0 is `Ready`. The Percona Server for MySQL cluster reports the `Initializing` state and the Operator sets the `AwaitingExternalBootstrap` status condition.

`mysql.autoRecovery`¶

Enables the Operator to recover the cluster automatically after a full cluster crash. Applies to clusters with Group Replication. If you omit the option, the Operator treats it as false. See Recover from a full cluster crash and the full-cluster-crash file.

Value type	Example
boolean	`true`

`mysql.vaultSecretName`¶

Specifies the secret for the HashiCorp Vault to carry on Data at Rest Encryption.

Value type	Example
string	`ps-cluster1-vault`

`mysql.size`¶

The number of the Percona Server for MySQL instances. This setting is required.

The safe range depends on mysql.clusterType. For async, use 2 or more instances (even counts are allowed). Size 1 requires unsafeFlags.mysqlSize. For group-replication, use an odd count between 3 and 9 unless you set that flag.

Value type	Example
int	`3`

`mysql.image`¶

The Docker image of the Percona Server for MySQL used (actual image names for Percona Server for MySQL 8.0 and Percona Server for MySQL 5.7 can be found in the list of certified images).

Value type	Example
string	`percona/percona-server:8.0.46-37.1`

`mysql.imagePullPolicy`¶

The policy used to update images .

Value type	Example
string	`Always`

`mysql.runtimeClassName`¶

Specifies the name of the RuntimeClass resource used to define and select the container runtime configuration.

Value type	Example
string	`image-rc`

`mysql.schedulerName`¶

The name of a Kubernetes scheduler used to assign MySQL Pods to Kubernetes nodes. The default-scheduler means kube-scheduler is used. You can define your custom schedulers here.

Value type	Example
string	`default-scheduler`

`mysql.priorityClassName`¶

The name of the Kubernetes PriorityClass , which is a way to assign priority levels to pods, helping the scheduler decide which pods to schedule first and which ones to evict last when resources are tight.

Value type	Example
string	`high-priority`

`mysql.nodeSelector`¶

Kubernetes nodeSelector .

Value type	Example
label	`disktype: ssd`

`mysql.serviceAccountName`¶

The Kubernetes Service Account for the MySQL Pods.

Value type	Example
string	`percona-server-mysql-operator-orchestrator`

`mysql.tolerations`¶

Specifies the Kubernetes tolerations applied to MySQL Pods allowing them to be scheduled on nodes with matching taints. Tolerations enable the Pod to tolerate specific node conditions, such as temporary unreachability or resource constraints, without being evicted immediately.

Value type	Example
subdoc	`node.alpha.kubernetes.io/unreachable`

`mysql.imagePullSecrets.name`¶

Specifies the Kubernetes imagePullSecrets for the MySQL image.

Value type	Example
string	`my-secret-1`

`mysql.initContainer.image`¶

An alternative init image for Percona Server for MySQL.

Value type	Example
string	`perconalab/percona-server-mysql-operator:1.3.0`

`mysql.initContainer.containerSecurityContext`¶

A custom Kubernetes Security Context for a Container used for the MySQL installation.

Value type	Example
subdoc	privileged: false runAsUser: 1001 runAsGroup: 1001

`mysql.initContainer.resources.requests.memory`¶

The Kubernetes memory requests for an image used for the MySQL installation.

Value type	Example
string	`100M`

`mysql.initContainer.resources.requests.cpu`¶

Kubernetes CPU requests for an image used for the MySQL installation.

Value type	Example
string	`100m`

`mysql.initContainer.resources.limits.memory`¶

Kubernetes memory limits for an image used for the MySQL installation.

Value type	Example
string	`200M`

`mysql.initContainer.resources.limits.cpu`¶

Kubernetes CPU limits for an image used for the MySQL installation.

Value type	Example
string	`200m`

`mysql.podDisruptionBudget.maxUnavailable`¶

The number of unavailable Pods your cluster can tolerate during voluntary disruption. It can be either an absolute value or a percentage.

To learn more, see podDisruptionBudgets .

Value type	Example
int	1

`mysql.podDisruptionBudget.minAvailable`¶

The number of Pods that must remain available during voluntary disruption. It can be either an absolute value or a percentage.

To learn more, see podDisruptionBudgets .

Value type	Example
int	0

`mysql.resources.requests.memory`¶

The Kubernetes memory requests for a Percona Server for MySQL container.

Value type	Example
string	`512M`

`mysql.resources.requests.cpu`¶

Kubernetes CPU requests for an image used for a Percona Server for MySQL container.

Value type	Example
string	`100m`

`mysql.resources.limits.memory`¶

Kubernetes memory limits for a Percona Server for MySQL container.

Value type	Example
string	`1G`

`mysql.resources.limits.cpu`¶

Kubernetes CPU limits for a Percona Server for MySQL container.

Value type	Example
string	`200m`

`mysql.startupProbe.initialDelaySeconds`¶

The number of seconds to wait before performing the startup probe .

Value type	Example
int	`15`

`mysql.startupProbe.timeoutSeconds`¶

The number of seconds after which the startup probe times out.

The startup probe runs MySQL bootstrap. On an asynchronous replica, bootstrap includes the clone, so this value limits how long that clone can run.

When spec.crVersion is 1.3.0 or higher, the default is 604800 seconds (7 days). A crVersion older than 1.3.0 keeps the previous default of 43200 seconds (12 hours). A value you set here replaces the default. See Set the clone stall timeout.

Value type	Example
int	`604800`

`mysql.startupProbe.periodSeconds`¶

How often to perform the startup probe . Measured in seconds.

Value type	Example
int	`10`

`mysql.startupProbe.successThreshold`¶

The number of successful probes required to mark the container successful.

Value type	Example
int	`1`

`mysql.startupProbe.failureThreshold`¶

The number of failed probes required to mark the container unready.

Value type	Example
int	`1`

`mysql.readinessProbe.initialDelaySeconds`¶

The number of seconds to wait before performing the first readiness probe .

Value type	Example
int	`30`

`mysql.readinessProbe.timeoutSeconds`¶

The number of seconds after which the readiness probe times out.

Value type	Example
int	`10`

`mysql.readinessProbe.periodSeconds`¶

How often to perform the readiness probe . Measured in seconds.

Value type	Example
int	`10`

`mysql.readinessProbe.successThreshold`¶

The number of successful probes required to mark the container successful.

Value type	Example
int	`1`

`mysql.readinessProbe.failureThreshold`¶

The number of failed probes required to mark the container unready.

Value type	Example
int	`3`

`mysql.livenessProbe.initialDelaySeconds`¶

The number of seconds to wait before performing the first liveness probe .

Value type	Example
int	`15`

`mysql.livenessProbe.timeoutSeconds`¶

The number of seconds after which the liveness probe times out.

Value type	Example
int	`10`

`mysql.livenessProbe.periodSeconds`¶

How often to perform the liveness probe . Measured in seconds.

Value type	Example
int	`10`

`mysql.livenessProbe.successThreshold`¶

The number of successful probes required to mark the container successful.

Value type	Example
int	`1`

`mysql.livenessProbe.failureThreshold`¶

The number of failed probes required to mark the container unhealthy.

Value type	Example
int	`3`

`mysql.env.name`¶

Name of an environment variable for MySQL Pods. The BOOTSTRAP_READ_TIMEOUT variable controls the timeout for bootstrapping the cluster. The BOOTSTRAP_CLONE_STALL_TIMEOUT sets how long an asynchronous clone can go without progress before bootstrap stops it for clusters with crVersion set to 1.3.0 or higher. See Set the clone stall timeout.

Read more about defining environment variables in Kubernetes documentation .

Value type	Example
string	`BOOTSTRAP_READ_TIMEOUT`

`mysql.env.value`¶

The value you set for the environment variables for a MySQL container.

Value type	Example
string	`"600"`

`mysql.envFrom.secretRef.name`¶

Name of a Secret or a ConfigMap, key/values of which are used as environment variables for MySQL Pods.

Value type	Example
string	`my-env-secret`

`mysql.affinity.antiAffinityTopologyKey`¶

The Operator topology key node anti-affinity constraint.

Value type	Example
string	`kubernetes.io/hostname`

`mysql.affinity.advanced`¶

In cases where the Pods require complex tuning the advanced option turns off the topologyKey effect. This setting allows the standard Kubernetes affinity constraints of any complexity to be used.

Value type	Example
subdoc

`mysql.topologySpreadConstraints.labelSelector.matchLabels`¶

The Label selector for the Kubernetes Pod Topology Spread Constraints .

Value type	Example
label	`app.kubernetes.io/name: percona-server`

`mysql.topologySpreadConstraints.maxSkew`¶

The degree to which Pods may be unevenly distributed under the Kubernetes Pod Topology Spread Constraints .

Value type	Example
int	1

`mysql.topologySpreadConstraints.topologyKey`¶

The key of node labels for the Kubernetes Pod Topology Spread Constraints .

Value type	Example
string	`kubernetes.io/hostname`

`mysql.topologySpreadConstraints.whenUnsatisfiable`¶

What to do with a Pod if it doesn’t satisfy the Kubernetes Pod Topology Spread Constraints .

Value type	Example
string	`DoNotSchedule`

`mysql.exposePrimary.enabled`¶

Enable or disable exposing Percona Server for MySQL primary node with dedicated IP addresses.

Value type	Example
boolean	`false`

`mysql.exposePrimary.type`¶

The Kubernetes Service Type used for exposure.

Value type	Example
string	`ClusterIP`

`mysql.exposePrimary.allocateLoadBalancerNodePorts`¶

Determines if Kubernetes should allocate NodePorts for the LoadBalancer Service. This option is relevant only when the Service type is set to LoadBalancer.

Value type	Example
boolean	`false`

`mysql.exposePrimary.annotations`¶

The Kubernetes annotations .

Value type	Example
string	`service.beta.kubernetes.io/aws-load-balancer-backend-protocol: tcp`, `service.beta.kubernetes.io/aws-load-balancer-type: nlb`

`mysql.exposePrimary.externalTrafficPolicy`¶

Specifies whether Service should route external traffic to cluster-wide (Cluster) or node-local (Local) endpoints; it can influence the load balancing effectiveness.

Value type	Example
string	`Cluster`

`mysql.exposePrimary.internalTrafficPolicy`¶

Specifies whether Service should route internal traffic to cluster-wide (Cluster) or node-local (Local) endpoints; it can influence the load balancing effectiveness.

Value type	Example
string	`Cluster`

`mysql.exposePrimary.labels`¶

Labels are key-value pairs attached to objects .

Value type	Example
label	`rack: rack-22`

`mysql.exposePrimary.loadBalancerSourceRanges`¶

The range of client IP addresses from which the load balancer should be reachable (if not set, there are no limitations).

Value type	Example
string	`10.0.0.0/8`

`mysql.expose.enabled`¶

Enable or disable exposing Percona Server for MySQL nodes with dedicated IP addresses. This setting exposes every Pod in your cluster.

Value type	Example
boolean	`true`

`mysql.expose.type`¶

The Kubernetes Service Type used for exposure.

Value type	Example
string	`ClusterIP`

`mysql.expose.allocateLoadBalancerNodePorts`¶

Determines if Kubernetes should allocate NodePorts for the LoadBalancer Service. This option is relevant only when the Service type is set to LoadBalancer.

Value type	Example
boolean	`false`

`mysql.expose.annotations`¶

The Kubernetes annotations .

Value type	Example
string	`service.beta.kubernetes.io/aws-load-balancer-backend-protocol: tcp`

`mysql.expose.externalTrafficPolicy`¶

Specifies whether Service should route external traffic to cluster-wide (Cluster) or node-local (Local) endpoints; it can influence the load balancing effectiveness.

Value type	Example
string	`Cluster`

`mysql.expose.internalTrafficPolicy`¶

Specifies whether Service should route internal traffic to cluster-wide (Cluster) or node-local (Local) endpoints; it can influence the load balancing effectiveness.

Value type	Example
string	`Cluster`

`mysql.expose.labels`¶

Labels are key-value pairs attached to objects .

Value type	Example
label	`rack: rack-22`

`mysql.expose.loadBalancerSourceRanges`¶

The range of client IP addresses from which the load balancer should be reachable (if not set, there are no limitations).

Value type	Example
string	`10.0.0.0/8`

`mysql.volumeSpec.emptyDir`¶

Starts a Pod with an empty temporary directory on the Kubernetes node. This directory exists as long as the Pod runs. Data is deleted when the Pod is deleted or moved to another node.

Value type	Example
string	`{}`

`mysql.volumeSpec.hostPath.path`¶

Specifies a path on the host node’s filesystem that will be mounted into your Pod when the Pod starts. Enables Pods to share files or access the host resources. Data persists as long as it exists on the host, independent of the Pod. Using the hostPath volume type presents many security risks.

Value type	Example
string	`/data`

`mysql.volumeSpec.hostPath.type`¶

Specifies a type for the hostPath volume.

Value type	Example
string	`Directory`

`mysql.volumeSpec.persistentVolumeClaim.storageClassName`¶

Requests a specific storage class for the Persistent Volume Claim. This will cause the PVC to match the right storage class if the cluster has StorageClasses enabled by the admin.

Value type	Example
string	`standard`

`mysql.volumeSpec.persistentVolumeClaim.accessModes`¶

Specify a specific access mode for a Persistent Volume. Kubernetes uses volume access modes to match PersistentVolumeClaims and PersistentVolumes.

Value type	Example
string	`"ReadWriteOnce"`

`mysql.volumeSpec.persistentVolumeClaim.resources.requests.storage`¶

The Kubernetes PersistentVolumeClaim size for the Percona Server for MySQL.

Value type	Example
string	`2Gi`

`mysql.gracePeriod`¶

Specifies the maximum time, in seconds, the Operator allows for a pod to shut down gracefully after receiving a termination signal before it is forcefully killed. This ensures critical cleanup tasks, like flushing data, can complete.

Value type	Example
int	`600`

`mysql.configuration`¶

The my.cnf file options to be passed to Percona Server for MySQL instances.

The Operator does not validate these options against your replication type. See Configuration compatibility considerations.

Value type	Example
string	`[mysqld]` `max_connections=250`

`mysql.autoConfig.enabled`¶

Enables automatic configuration tuning. The Operator calculates a full MySQL configuration based on the Pod resources, database workload profile, and MySQL version. Requires mysql.resources (CPU and memory) and mysql.autoConfig.version to be set.

Value type	Example
boolean	`true`

`mysql.autoConfig.loadType`¶

The workload profile used to calculate the configuration. If not set, the Operator defaults to someWrites.

See Choose a workload profile for what each value means.

Value type	Allowed values	Example
string	`mostlyReads`, `someWrites`, `equalReadsWrites`, `heavyWrites`	`someWrites`

`mysql.autoConfig.version`¶

The MySQL version the configuration is calculated for, in the <major>.<minor>.<patch> format. This option controls the parameters set for the selected Percona Server for MySQL version. The mysql.image option controls the image version to be deployed. The Operator reads these two values separately. You must keep mysql.autoConfig.version and mysql.image values in sync yourself. Read more in the Keep the MySQL version in sync.

Value type	Example
string	`8.4.8`

`mysql.sidecars.image`¶

Image for the custom sidecar container for Percona Server for MySQL Pods.

Value type	Example
string	`busybox`

`mysql.sidecars.command`¶

Command for the custom sidecar container for Percona Server for MySQL Pods.

Value type	Example
array	`["sleep", "30d"]`

`mysql.sidecars.name`¶

Name of the custom sidecar container for Percona Server for MySQL Pods.

Value type	Example
string	`my-sidecar-1`

`mysql.sidecars.volumeMounts.mountPath`¶

Mount path of the custom sidecar container volume for Replica Set Pods.

Value type	Example
string	`/volume1`

`mysql.sidecars.resources.requests.memory`¶

The Kubernetes memory requests for a Percona Server for MySQL sidecar container.

Value type	Example
string	`16M`

`mysql.sidecars.volumeMounts.name`¶

Name of the custom sidecar container volume for Replica Set Pods.

Value type	Example
string	`sidecar-volume-claim`

`mysql.sidecarVolumes`¶

Volume specification for the custom sidecar container volume for Percona Server for MySQL Pods.

Value type	Example
subdoc

`mysql.sidecarPVCs`¶

Persistent Volume Claim for the custom sidecar container volume for Replica Set Pods |

Value type	Example
subdoc

HAProxy subsection¶

The proxy.haproxy subsection in the deploy/cr.yaml file contains configuration options for the HAProxy service.

`proxy.haproxy.enabled`¶

Enables or disables load balancing with HAProxy Services .

Value type	Example
boolean	`true`

`proxy.haproxy.size`¶

The number of the HAProxy Pods to provide load balancing. Safe configuration should have 2 or more. This setting is required.

Value type	Example
int	`3`

`proxy.haproxy.image`¶

HAProxy Docker image to use.

Value type	Example
string	`percona/perconalab-xtradb-cluster-operator:1.3.0-haproxy`

`proxy.haproxy.imagePullPolicy`¶

The policy used to update images .

Value type	Example
string	`Always`

`proxy.haproxy.schedulerName`¶

The name of a Kubernetes scheduler used to assign HAProxy Pods to Kubernetes nodes. The default-scheduler means kube-scheduler is used. You can define your custom schedulers here.

Value type	Example
string	`default-scheduler`

`proxy.haproxy.priorityClassName`¶

The name of the Kubernetes PriorityClass , which is a way to assign priority levels to pods, helping the scheduler decide which pods to schedule first and which ones to evict last when resources are tight.

Value type	Example
string	`high-priority`

`proxy.haproxy.nodeSelector`¶

Kubernetes nodeSelector . It enables you to define node labels thereby ensuring that Pods will be scheduled onto nodes that have each of the labels you specify.

Value type	Example
label	`disktype: ssd`

`proxy.haproxy.serviceAccountName`¶

The Kubernetes Service Account for the HAProxy Pods.

Value type	Example
string	`percona-server-mysql-operator-orchestrator`

`proxy.haproxy.podDisruptionBudget.maxUnavailable`¶

The number of unavailable Pods your cluster can tolerate during voluntary disruption. It can be either an absolute value or a percentage.

To learn more, see podDisruptionBudgets .

Value type	Example
int	1

`proxy.haproxy.podDisruptionBudget.minAvailable`¶

The number of Pods that must remain available during voluntary disruption. It can be either an absolute value or a percentage.

To learn more, see podDisruptionBudgets .

Value type	Example
int	0

`proxy.haproxy.runtimeClassName`¶

Specifies the name of the RuntimeClass resource used to define and select the container runtime configuration.

Value type	Example
string	`image-rc`

`proxy.haproxy.tolerations`¶

Specifies the Kubernetes tolerations applied to HAProxy Pods allowing them to be scheduled on nodes with matching taints. Tolerations enable the Pod to tolerate specific node conditions, such as resource constraints being temporary unreachable, without being evicted immediately.

Value type	Example
subdoc	`node.alpha.kubernetes.io/unreachable`

`proxy.haproxy.imagePullSecrets.name`¶

Specifies the Kubernetes imagePullSecrets for the HAProxy image.

Value type	Example
string	`my-secret-1`

`proxy.haproxy.resources.requests.memory`¶

The Kubernetes memory requests for the main HAProxy container.

Value type	Example
string	`1G`

`proxy.haproxy.resources.requests.cpu`¶

Kubernetes CPU requests for the main HAProxy container.

Value type	Example
string	`600m`

`proxy.haproxy.resources.limits.memory`¶

Kubernetes memory limits for the main HAProxy container.

Value type	Example
string	`1G`

`proxy.haproxy.resources.limits.cpu`¶

Kubernetes CPU limits for the main HAProxy container.

Value type	Example
string	`700m`

`proxy.haproxy.sidecarResources.CONTAINER-NAME.requests.memory`¶

The Kubernetes memory requests for an HAProxy sidecar container (for example, mysql-monit).

Value type	Example
string	`64Mi`

`proxy.haproxy.sidecarResources.CONTAINER-NAME.requests.cpu`¶

Kubernetes CPU requests for an HAProxy sidecar container (for example, mysql-monit).

Value type	Example
string	`50m`

`proxy.haproxy.sidecarResources.CONTAINER-NAME.limits.memory`¶

Kubernetes memory limits for an HAProxy sidecar container (for example, mysql-monit).

Value type	Example
string	`128Mi`

`proxy.haproxy.sidecarResources.CONTAINER-NAME.limits.cpu`¶

Kubernetes CPU limits for an HAProxy sidecar container (for example, mysql-monit).

Value type	Example
string	`100m`

`proxy.haproxy.env.name`¶

Name of an environment variable for HAProxy.

Value type	Example
string	`HA_CONNECTION_TIMEOUT`

`proxy.haproxy.env.value`¶

Value of an environment variable for HAProxy.

Value type	Example
string	`"1000"`

`proxy.haproxy.envFrom.secretRef.name`¶

Name of a Secret with environment variables for HAProxy.

Value type	Example
string	`haproxy-env-secret`

`proxy.haproxy.startupProbe.initialDelaySeconds`¶

The number of seconds to wait before performing the startup probe .

Value type	Example
int	`15`

`proxy.haproxy.startupProbe.timeoutSeconds`¶

The number of seconds after which the startup probe times out.

Value type	Example
int	`43200`

`proxy.haproxy.startupProbe.periodSeconds`¶

How often to perform the startup probe . Measured in seconds.

Value type	Example
int	`10`

`proxy.haproxy.startupProbe.successThreshold`¶

The number of successful probes required to mark the container successful.

Value type	Example
int	`1`

`proxy.haproxy.startupProbe.failureThreshold`¶

The number of failed probes required to mark the container unready.

Value type	Example
int	`1`

`proxy.haproxy.readinessProbe.timeoutSeconds`¶

Number of seconds after which the readiness probe times out.

Value type	Example
int	`3`

`proxy.haproxy.readinessProbe.periodSeconds`¶

How often (in seconds) to perform the readiness probe .

Value type	Example
int	`5`

`proxy.haproxy.readinessProbe.successThreshold`¶

Minimum consecutive successes for the readiness probe to be considered successful after having failed.

Value type	Example
int	`3`

`proxy.haproxy.readinessProbe.failureThreshold`¶

When the readiness probe fails, Kubernetes will try this number of times before marking the Pod Unready.

Value type	Example
int	`1`

`proxy.haproxy.livenessProbe.timeoutSeconds`¶

Number of seconds after which the liveness probe times out.

Value type	Example
int	`3`

`proxy.haproxy.livenessProbe.periodSeconds`¶

How often (in seconds) to perform the liveness probe .

Value type	Example
int	`5`

`proxy.haproxy.livenessProbe.successThreshold`¶

Minimum consecutive successes for the liveness probe to be considered successful after having failed.

Value type	Example
int	`3`

`proxy.haproxy.livenessProbe.failureThreshold`¶

When the liveness probe fails, Kubernetes will try this number of times before marking the Pod Unready.

Value type	Example
int	`1`

`proxy.haproxy.gracePeriod`¶

Specifies the maximum time, in seconds, the Operator allows for a pod to shut down gracefully after receiving a termination signal before it is forcefully killed. This ensures critical cleanup tasks, like flushing data, can complete.

Value type	Example
int	`30`

`proxy.haproxy.configuration`¶

The custom HAProxy configuration file contents.

Value type	Example
string

`proxy.haproxy.antiAffinityTopologyKey`¶

The Operator topology key node anti-affinity constraint.

Value type	Example
string	`kubernetes.io/hostname`

`proxy.haproxy.affinity.advanced`¶

If available it makes a topologyKey node affinity constraint to be ignored.

Value type	Example
subdoc

`proxy.haproxy.topologySpreadConstraints.labelSelector.matchLabels`¶

The Label selector for the Kubernetes Pod Topology Spread Constraints .

Value type	Example
label	`app.kubernetes.io/name: percona-server`

`proxy.haproxy.topologySpreadConstraints.maxSkew`¶

The degree to which Pods may be unevenly distributed under the Kubernetes Pod Topology Spread Constraints .

Value type	Example
int	1

`proxy.haproxy.topologySpreadConstraints.topologyKey`¶

The key of node labels for the Kubernetes Pod Topology Spread Constraints .

Value type	Example
string	`kubernetes.io/hostname`

`proxy.haproxy.topologySpreadConstraints.whenUnsatisfiable`¶

What to do with a Pod if it doesn’t satisfy the Kubernetes Pod Topology Spread Constraints .

Value type	Example
string	`DoNotSchedule`

`proxy.haproxy.expose.type`¶

The Kubernetes Service Type used for HAProxy exposure.

Value type	Example
string	`ClusterIP`

`proxy.haproxy.expose.allocateLoadBalancerNodePorts`¶

Determines if Kubernetes should allocate NodePorts for the LoadBalancer Service. This option is relevant only when the Service type is set to LoadBalancer.

Value type	Example
boolean	`false`

`proxy.haproxy.expose.annotations`¶

The Kubernetes annotations for HAProxy.

Value type	Example
string	`service.beta.kubernetes.io/aws-load-balancer-backend-protocol: tcp`

`proxy.haproxy.expose.externalTrafficPolicy`¶

Specifies whether Service for HAProxy should route external traffic to cluster-wide (Cluster) or node-local (Local) endpoints; it can influence the load balancing effectiveness.

Value type	Example
string	`Cluster`

`proxy.haproxy.expose.internalTrafficPolicy`¶

Specifies whether Service for HAProxy should route internal traffic to cluster-wide (Cluster) or node-local (Local) endpoints; it can influence the load balancing effectiveness.

Value type	Example
string	`Cluster`

`proxy.haproxy.expose.labels`¶

Labels are key-value pairs attached to objects for HAProxy.

Value type	Example
label	`rack: rack-22`

`proxy.haproxy.expose.loadBalancerIP`¶

The static IP-address for the load balancer. This field is deprecated in Kubernetes 1.24+ and removed from the Operator starting with version 0.10.0. If you have defined it, refer to the Kubernetes documentation for recommendations.

Value type	Example
string	`127.0.0.1`

`proxy.haproxy.expose.loadBalancerSourceRanges`¶

The range of client IP addresses from which the load balancer should be reachable (if not set, there is no limitations) |

Value type	Example
string	`10.0.0.0/8`

`proxy.haproxy.containerSecurityContext`¶

A custom Kubernetes Security Context for a Container used for the HAProxy installation.

Value type	Example
subdoc	privileged: false runAsUser: 1001 runAsGroup: 1001

`proxy.haproxy.podSecurityContext`¶

A custom Kubernetes Security Context for a Pod to be used instead of the default one.

Value type	Example
subdoc	fsGroup: 1001 supplementalGroups: [1001, 1002, 1003]

Router subsection¶

The proxy.router subsection in the deploy/cr.yaml file contains configuration options for the MySQL Router , which can act as a proxy for Group replication.

`proxy.router.enabled`¶

Enables or disables MySQL Router.

Value type	Example
boolean	`false`

`proxy.router.size`¶

The number of the Router Pods to provide routing to MySQL Servers. This setting is required.

Value type	Example
int	`3`

`proxy.router.image`¶

Router Docker image to use.

Value type	Example
string	`perconalab/percona-server-mysql-operator:1.3.0-router`

`proxy.router.imagePullPolicy`¶

The policy used to update images .

Value type	Example
string	`Always`

`proxy.router.runtimeClassName`¶

Specifies the name of the RuntimeClass resource used to define and select the container runtime configuration.

Value type	Example
string	`image-rc`

`proxy.router.tolerations`¶

Specifies the Kubernetes tolerations applied to Router Pods allowing them to be scheduled on nodes with matching taints. Tolerations enable the Pod to tolerate specific node conditions, such as temporary unreachability or resource constraints, without being evicted immediately.

Value type	Example
subdoc	`node.alpha.kubernetes.io/unreachable`

`proxy.router.imagePullSecrets.name`¶

Specifies the Kubernetes imagePullSecrets for the Router image.

Value type	Example
string	`my-secret-1`

`proxy.router.initContainer.image`¶

An alternative init image for MySQL Router Pods.

Value type	Example
string	`perconalab/percona-server-mysql-operator:1.3.0`

`proxy.router.initContainer.containerSecurityContext`¶

A custom Kubernetes Security Context for a Container for the image used for MySQL Router Pods installation.

Value type	Example
subdoc	privileged: false runAsUser: 1001 runAsGroup: 1001

`proxy.router.initContainer.resources.requests.memory`¶

The Kubernetes memory requests for an image used for MySQL Router Pods installation.

Value type	Example
string	`100M`

`proxy.router.initContainer.resources.requests.cpu`¶

Kubernetes CPU requests for an image used for MySQL Router Pods installation.

Value type	Example
string	`100m`

`proxy.router.initContainer.resources.limits.memory`¶

Kubernetes memory limits for an image used for MySQL Router Pods installation.

Value type	Example
string	`200M`

`proxy.router.initContainer.resources.limits.cpu`¶

Kubernetes CPU limits for an image used for MySQL Router Pods installation.

Value type	Example
string	`200m`

`proxy.router.env.name`¶

Name of an environment variable for MySQL Router Pods. Read more about defining environment variables in Kubernetes documentation .

Value type	Example
string	`MY_ENV`

`proxy.router.env.value`¶

Value of an environment variable for MySQL Router Pods.

Value type	Example
string	`"1000"`

`proxy.router.envFrom.secretRef.name`¶

Name of a Secret or a ConfigMap, key/values of which are used as environment variables for MySQL Router Pods.

Value type	Example
string	`my-env-secret`

`proxy.router.podDisruptionBudget.maxUnavailable`¶

The number of unavailable Pods your cluster can tolerate during voluntary disruption. It can be either an absolute value or a percentage.

To learn more, see podDisruptionBudgets .

Value type	Example
int	1

`proxy.router.podDisruptionBudget.minAvailable`¶

The number of Pods that must remain available during voluntary disruption. It can be either an absolute value or a percentage.

To learn more, see podDisruptionBudgets .

Value type	Example
int	0

`proxy.router.ports.name`¶

The name for a custom or an existing port for the MySQL Router Service.

Value type	Example
string	`http`

`proxy.router.ports.port`¶

The port exposed by the MySQL Router service to the outside world or other components.

Value type	Example
string	`8443`

`proxy.router.ports.targetPort`¶

The port inside the Pod/container where MySQL Router is actually listening. When a client connects to the external port, Kubernetes forwards that traffic to the targetPort value on the backend Pod. A zero (0) value means Kubernetes uses the default internal port or does not do the remapping.

Value type	Example
string	`0`

`proxy.router.resources.requests.memory`¶

The Kubernetes memory requests for MySQL Router container.

Value type	Example
string	`256M`

`proxy.router.resources.requests.cpu`¶

Kubernetes CPU requests for MySQL Router container.

Value type	Example
string	`100m`

`proxy.router.resources.limits.memory`¶

Kubernetes memory limits for MySQL Router container.

Value type	Example
string	`256M`

`proxy.router.resources.limits.cpu`¶

Kubernetes CPU limits for MySQL Router container.

Value type	Example
string	`200m`

`proxy.router.affinity.antiAffinityTopologyKey`¶

The Operator topology key node anti-affinity constraint.

Value type	Example
string	`kubernetes.io/hostname`

`proxy.router.affinity.advanced`¶

In cases where the Pods require complex tuning the advanced option turns off the topologyKey effect. This setting allows the standard Kubernetes affinity constraints of any complexity to be used.

Value type	Example
subdoc

`proxy.router.topologySpreadConstraints.labelSelector.matchLabels`¶

The Label selector for the Kubernetes Pod Topology Spread Constraints .

Value type	Example
label	`app.kubernetes.io/name: percona-server`

`proxy.router.topologySpreadConstraints.maxSkew`¶

The degree to which Pods may be unevenly distributed under the Kubernetes Pod Topology Spread Constraints .

Value type	Example
int	1

`proxy.router.topologySpreadConstraints.topologyKey`¶

The key of node labels for the Kubernetes Pod Topology Spread Constraints .

Value type	Example
string	`kubernetes.io/hostname`

`proxy.router.topologySpreadConstraints.whenUnsatisfiable`¶

What to do with a Pod if it doesn’t satisfy the Kubernetes Pod Topology Spread Constraints .

Value type	Example
string	`DoNotSchedule`

`proxy.router.gracePeriod`¶

Specifies the maximum time, in seconds, the Operator allows for a pod to shut down gracefully after receiving a termination signal before it is forcefully killed. This ensures critical cleanup tasks, like flushing data, can complete.

Value type	Example
int	`30`

`proxy.router.configuration`¶

Custom configuration options to be passed to MySQL Router.

Value type	Example
string	\| `[default]` `logging_folder=/tmp/router/log` `[logger]` `level=DEBUG`

`proxy.router.expose.type`¶

The Kubernetes Service Type used for MySQL Router instances exposure.

Value type	Example
string	`ClusterIP`

`proxy.router.expose.allocateLoadBalancerNodePorts`¶

Determines if Kubernetes should allocate NodePorts for the LoadBalancer Service. This option is relevant only when the Service type is set to LoadBalancer.

Value type	Example
boolean	`false`

`proxy.router.expose.annotations`¶

The Kubernetes annotations for MySQL Router.

Value type	Example
string	`service.beta.kubernetes.io/aws-load-balancer-backend-protocol: tcp`

`proxy.router.expose.externalTrafficPolicy`¶

Specifies whether Service for MySQL Router should route external traffic to cluster-wide (Cluster) or node-local (Local) endpoints; it can influence the load balancing effectiveness.

Value type	Example
string	`Cluster`

`proxy.router.expose.internalTrafficPolicy`¶

Specifies whether Service for MySQL Router should route internal traffic to cluster-wide (Cluster) or node-local (Local) endpoints; it can influence the load balancing effectiveness.

Value type	Example
string	`Cluster`

`proxy.router.expose.labels`¶

Labels are key-value pairs attached to objects for MySQL Router.

Value type	Example
label	`rack: rack-22`

`proxy.router.expose.loadBalancerIP`¶

The static IP-address for the load balancer. This field is deprecated in Kubernetes 1.24+ and removed from the Operator starting with version 0.10.0. If you have defined it, refer to the Kubernetes documentation for recommendations.

Value type	Example
string	`127.0.0.1`

`proxy.router.expose.loadBalancerSourceRanges`¶

The range of client IP addresses from which the load balancer should be reachable (if not set, there is no limitations) |

Value type	Example
string	`10.0.0.0/8`

Orchestrator section¶

The orchestrator section in the deploy/cr.yaml file contains configuration options for the Orchestrator - a replication topology manager, used if asynchronous replication is turned on.

`orchestrator.enabled`¶

Enables or disables the Orchestrator.

Orchestrator is used with asynchronous replication. Disable it when you switch from async to group replication. Enable it when you switch from group replication to async.

Value type	Example
boolean	`true`

`orchestrator.failover.timeout`¶

How long the cluster tries to recover the transactions from the failed primary’s binlog. Default is 1h.

The timeout is measured from the moment Orchestrator first detects the failure and covers every retry, not a single attempt. If the Orchestrator Raft leader changes while a failover is running, the timeout restarts. This delays the decision what to do when the timeout expires.

Only applies to async clusters, and only takes effect starting with crVersion 1.3.0. With an older crVersion, the Operator validates this field but ignores it. See Understand failover timeout and recovery policy.

Value type	Example
string	`1h`

`orchestrator.failover.onTimeout`¶

What the Operator does once orchestrator.failover.timeout expires without recovering the transactions stranded on a failed primary:

Abort (default) — leaves the cluster without a writable primary until you intervene. No transactions are lost, but writes stay blocked. Promote a replica yourself with the percona.com/force-promote-with-possible-data-loss annotation.
ForceWithPossibleDataLoss — promotes the best available candidate anyway. Transactions the old primary committed and never replicated are lost.

Only applies to async clusters, and only takes effect starting with crVersion 1.3.0. See Understand failover timeout and recovery policy.

Value type	Example
string	`Abort`

`orchestrator.size`¶

The number of the Orchestrator Pods to provide load balancing. This setting is required.

Value type	Example
int	`3`

`orchestrator.image`¶

Orchestrator Docker image to use.

Value type	Example
string	`perconalab/percona-server-mysql-operator:1.3.0-orchestrator`

`orchestrator.imagePullPolicy`¶

The policy used to update images .

Value type	Example
string	`Always`

`orchestrator.runtimeClassName`¶

Specifies the name of the RuntimeClass resource used to define and select the container runtime configuration.

Value type	Example
string	`image-rc`

`orchestrator.schedulerName`¶

The name of a Kubernetes scheduler used to assign Orchestrator Pods to Kubernetes nodes. The default-scheduler means kube-scheduler is used. You can define your custom schedulers here.

Value type	Example
string	`default-scheduler`

`orchestrator.priorityClassName`¶

The name of the Kubernetes PriorityClass , which is a way to assign priority levels to pods, helping the scheduler decide which pods to schedule first and which ones to evict last when resources are tight.

Value type	Example
string	`high-priority`

`orchestrator.nodeSelector`¶

Kubernetes nodeSelector . It enables you to define node labels thereby ensuring that Pods will be scheduled onto nodes that have each of the labels you specify.

Value type	Example
label	`disktype: ssd`

`orchestrator.startupProbe.initialDelaySeconds`¶

The number of seconds to wait before performing the startup probe .

Value type	Example
int	`15`

`orchestrator.startupProbe.timeoutSeconds`¶

The number of seconds after which the startup probe times out.

Value type	Example
int	`43200`

`orchestrator.startupProbe.periodSeconds`¶

How often to perform the startup probe . Measured in seconds.

Value type	Example
int	`10`

`orchestrator.startupProbe.successThreshold`¶

The number of successful probes required to mark the container successful.

Value type	Example
int	`1`

`orchestrator.startupProbe.failureThreshold`¶

The number of failed probes required to mark the container unready.

Value type	Example
int	`1`

`orchestrator.readinessProbe.timeoutSeconds`¶

Number of seconds after which the readiness probe times out.

Value type	Example
int	`3`

`orchestrator.readinessProbe.periodSeconds`¶

How often (in seconds) to perform the readiness probe .

Value type	Example
int	`5`

`orchestrator.readinessProbe.successThreshold`¶

Minimum consecutive successes for the readiness probe to be considered successful after having failed.

Value type	Example
int	`3`

`orchestrator.readinessProbe.failureThreshold`¶

When the readiness probe fails, Kubernetes will try this number of times before marking the Pod Unready.

Value type	Example
int	`1`

`orchestrator.livenessProbe.timeoutSeconds`¶

Number of seconds after which the liveness probe times out.

Value type	Example
int	`3`

`orchestrator.livenessProbe.periodSeconds`¶

How often (in seconds) to perform the liveness probe .

Value type	Example
int	`5`

`orchestrator.livenessProbe.successThreshold`¶

Minimum consecutive successes for the liveness probe to be considered successful after having failed.

Value type	Example
int	`3`

`orchestrator.livenessProbe.failureThreshold`¶

When the liveness probe fails, Kubernetes will try this number of times before marking the Pod Unready.

Value type	Example
int	`1`

`orchestrator.env.name`¶

Name of an environment variable for Orchestrator Pods. Read more about defining environment variables in Kubernetes documentation .

Value type	Example
string	`MY_ENV`

`orchestrator.env.value`¶

Value of an environment variable for Orchestrator Pods.

Value type	Example
string	`"1000"`

`orchestrator.envFrom.secretRef.name`¶

Name of a Secret or a ConfigMap, key/values of which are used as environment variables for Orchestrator Pods.

Value type	Example
string	`my-env-secret`

`orchestrator.tolerations`¶

Specifies the Kubernetes tolerations applied to Orchestrator Pods allowing them to be scheduled on nodes with matching taints. Tolerations enable the Pod to tolerate specific node conditions, such as temporary unreachability or resource constraints, without being evicted immediately.

Value type	Example
subdoc	`node.alpha.kubernetes.io/unreachable`

`orchestrator.imagePullSecrets.name`¶

Specifies the Kubernetes imagePullSecrets for the Orchestrator image.

Value type	Example
string	`my-secret-1`

`orchestrator.serviceAccountName`¶

The Kubernetes Service Account for the Orchestrator Pods.

Value type	Example
string	`percona-server-mysql-operator-orchestrator`

`orchestrator.initContainer.image`¶

An alternative init image for Orchestrator Pods.

Value type	Example
string	`perconalab/percona-server-mysql-operator:1.3.0`

`orchestrator.initContainer.containerSecurityContext`¶

A custom Kubernetes Security Context for a Container for the image used for Orchestrator Pods installation.

Value type	Example
subdoc	privileged: false runAsUser: 1001 runAsGroup: 1001

`orchestrator.initContainer.resources.requests.memory`¶

The Kubernetes memory requests for an image used for Orchestrator Pods installation.

Value type	Example
string	`100M`

`orchestrator.initContainer.resources.requests.cpu`¶

Kubernetes CPU requests for an image used for Orchestrator Pods installation.

Value type	Example
string	`100m`

`orchestrator.initContainer.resources.limits.memory`¶

Kubernetes memory limits for an image used for Orchestrator Pods installation.

Value type	Example
string	`200M`

`orchestrator.initContainer.resources.limits.cpu`¶

Kubernetes CPU limits for an image used for Orchestrator Pods installation.

Value type	Example
string	`200m`

`orchestrator.affinity.antiAffinityTopologyKey`¶

The Operator topology key node anti-affinity constraint.

Value type	Example
string	`kubernetes.io/hostname`

`orchestrator.affinity.advanced`¶

In cases where the Pods require complex tuning the advanced option turns off the topologyKey effect. This setting allows the standard Kubernetes affinity constraints of any complexity to be used.

Value type	Example
subdoc

`orchestrator.topologySpreadConstraints.labelSelector.matchLabels`¶

The Label selector for the Kubernetes Pod Topology Spread Constraints .

Value type	Example
label	`app.kubernetes.io/name: percona-server`

`orchestrator.topologySpreadConstraints.maxSkew`¶

The degree to which Pods may be unevenly distributed under the Kubernetes Pod Topology Spread Constraints .

Value type	Example
int	1

`orchestrator.topologySpreadConstraints.topologyKey`¶

The key of node labels for the Kubernetes Pod Topology Spread Constraints .

Value type	Example
string	`kubernetes.io/hostname`

`orchestrator.topologySpreadConstraints.whenUnsatisfiable`¶

What to do with a Pod if it doesn’t satisfy the Kubernetes Pod Topology Spread Constraints .

Value type	Example
string	`DoNotSchedule`

`orchestrator.gracePeriod`¶

Specifies the maximum time, in seconds, the Operator allows for a pod to shut down gracefully after receiving a termination signal before it is forcefully killed. This ensures critical cleanup tasks, like flushing data, can complete.

Value type	Example
int	`30`

`orchestrator.expose.type`¶

The Kubernetes Service Type used to expose the Orchestrator HTTP API (port 3000). The default ClusterIP keeps the API reachable only inside the cluster.

When you change this option to LoadBalancer or NodePort, the Orchestrator API becomes reachable from outside the cluster. Starting with Operator version 1.2.0, when spec.crVersion is 1.2.0 or higher, all API requests must authenticate. To learn more about authentication, refer to the Orchestrator HTTP API authentication section.

Value type	Example
string	`ClusterIP`

`orchestrator.expose.allocateLoadBalancerNodePorts`¶

Determines if Kubernetes should allocate NodePorts for the LoadBalancer Service. This option is relevant only when the Service type is set to LoadBalancer.

Value type	Example
boolean	`false`

`orchestrator.expose.annotations`¶

The Kubernetes annotations for the Orchestrator.

Value type	Example
string	`service.beta.kubernetes.io/aws-load-balancer-backend-protocol: tcp`

`orchestrator.expose.externalTrafficPolicy`¶

Specifies whether Service for the Orchestrator should route external traffic to cluster-wide (Cluster) or node-local (Local) endpoints; it can influence the load balancing effectiveness.

Value type	Example
string	`Cluster`

`orchestrator.expose.internalTrafficPolicy`¶

Specifies whether Service for the Orchestrator should route internal traffic to cluster-wide (Cluster) or node-local (Local) endpoints; it can influence the load balancing effectiveness.

Value type	Example
string	`Cluster`

`orchestrator.expose.labels`¶

Labels are key-value pairs attached to objects for the Orchestrator.

Value type	Example
label	`rack: rack-22`

`orchestrator.expose.loadBalancerSourceRanges`¶

The range of client IP addresses from which the load balancer should be reachable (if not set, there is no limitations).

Value type	Example
string	`10.0.0.0/8`

`orchestrator.containerSecurityContext`¶

A custom Kubernetes Security Context for a Container used for the Orchestrator installation.

Value type	Example
subdoc	privileged: false runAsUser: 1001 runAsGroup: 1001

`orchestrator.podSecurityContext`¶

A custom Kubernetes Security Context for a Pod to be used instead of the default one.

Value type	Example
subdoc	fsGroup: 1001 supplementalGroups: [1001, 1002, 1003]

`orchestrator.podDisruptionBudget.maxUnavailable`¶

The number of unavailable Pods your cluster can tolerate during voluntary disruption. It can be either an absolute value or a percentage.

To learn more, see podDisruptionBudgets .

Value type	Example
int	1

`orchestrator.podDisruptionBudget.minAvailable`¶

The number of Pods that must remain available during voluntary disruption. It can be either an absolute value or a percentage.

To learn more, see podDisruptionBudgets .

Value type	Example
int	0

`orchestrator.resources.requests.memory`¶

The Kubernetes memory requests for an Orchestrator container.

Value type	Example
string	`128M`

`orchestrator.resources.requests.cpu`¶

Kubernetes CPU requests for an Orchestrator container.

Value type	Example
string	`100m`

`orchestrator.resources.limits.memory`¶

Kubernetes memory limits for an Orchestrator container.

Value type	Example
string	`256M`

`orchestrator.resources.limits.cpu`¶

Kubernetes CPU limits for an Orchestrator container.

Value type	Example
string	`200m`

`orchestrator.volumeSpec.persistentVolumeClaim.resources.requests.storage`¶

The Kubernetes PersistentVolumeClaim size for the Orchestrator |

Value type	Example
string	`1Gi`

`orchestrator.configuration`¶

Custom Orchestrator options to merge into the Orchestrator configuration. The value must be a JSON object passed as a string. You cannot override keys managed by the Operator, such as Raft topology, topology TLS, HTTP authentication, failover hooks, alias detection queries, and storage paths. Changing them is silently ignored.

Starting with Custom Resource version 1.3.0, the Operator sets ReasonableReplicationLagSeconds to 60 by default. Orchestrator reports a replica that lags behind the primary by more than this value, and the Operator reflects it in the ReplicationLagging condition. You can override this value here.

For an async cluster with crVersion 1.3.0 or later, the Operator also derives and pins the settings the zero-data-loss failover mechanism depends on, and ignores any override of them here:

Orchestrator key	Derived from
`FailMasterPromotionOnLagMinutes`	Always `0`
`DelayMasterPromotionIfSQLThreadNotUpToDate`, `FailMasterPromotionIfSQLThreadNotUpToDate`	Always `false`
`PreFailoverProcesses`, `PostFailoverProcesses`, `PostUnsuccessfulFailoverProcesses`	The failover mechanism itself
`RecoveryPeriodBlockSeconds`	`orchestrator.failover.timeout` + 1 hour
`UnseenInstanceForgetHours`	`orchestrator.failover.timeout` + 1 hour, rounded up

The following setting is not pinned, and you can override it:

ReasonableMaintenanceReplicationLagSeconds is the maximum replication lag, in seconds, that a candidate replica can have for Orchestrator to allow a planned switchover. Orchestrator rejects any replica that lags more. After the primary becomes read-only, Orchestrator also waits at most this long for the candidate to catch up. With crVersion 1.3.0 or later, the default is 120, twice ReasonableReplicationLagSeconds, which is 60 by default. With an older crVersion, the value from the Orchestrator image applies, which is 300. See Set the replication lag limit for a planned switchover.

Updating this field triggers a rolling restart of the Orchestrator Pods.

Value type	Example
string	`'{"ReplicationLagQuery": "..."}'`

Users section¶

The users section in the deploy/cr.yaml file contains configuration options to create custom MySQL users via the Custom Resource.

`users.name`¶

The username of the MySQL user. This field is required.

Value type	Example
string	`my-user`

`users.dbs`¶

Databases that the user can access. If a specified database does not exist, the Operator creates it. When you omit this field and set grants, privileges apply to all databases (*.*). Omit this field when you set administrative (global) grants such as SHUTDOWN, because those privileges apply at the global level.

Value type	Example
array	`["db1", "db2"]`

`users.hosts`¶

Hosts that the user can connect from. If not specified, defaults to '%', enabling the user to connect from any host.

Value type	Example
array	`["localhost"]`

`users.passwordSecretRef.name`¶

Name of the Secret that contains the user’s password. If not provided, the Operator generates a password and stores it in a Secret named <cluster-name>-user-<username> (for example, ps-cluster1-user-my-user).

Value type	Example
string	`my-user-pwd`

`users.passwordSecretRef.key`¶

Key in the Secret that holds the user’s password (password by default).

Value type	Example
string	`password`

`users.withGrantOption`¶

When true, the user can grant their own privileges to other users. The default is false.

Value type	Example
boolean	`false`

`users.grants`¶

Privileges granted to the user. If omitted, the Operator creates the user without additional GRANT statements.

Value type	Example
array	`["SELECT", "DELETE", "INSERT"]`

PMM section¶

The pmm section in the deploy/cr.yaml file contains configuration options for Percona Monitoring and Management.

`pmm.enabled`¶

Enables or disables monitoring Percona Server for MySQL with PMM.

Value type	Example
boolean	`false`

`pmm.image`¶

PMM client Docker image to use.

Value type	Example
string	`percona/pmm-client:3.10.0`

`pmm.imagePullPolicy`¶

The policy used to update images .

Value type	Example
string	`Always`

`pmm.mysqlParams`¶

Enables to pass MySQL parameters to PMM. For example, to change the number of tables (from the default of 1000) beyond which per-table statistics collection is disabled.

Value type	Example
string	`"--disable-tablestats-limit=2000"`

`pmm.haproxyParams`¶

Additional parameters passed to the pmm-admin add haproxy command for HAProxy Pods. Use this to add custom labels, skip the connection check, or change the metrics listen port.

See the PMM documentation for the list of available flags.

Note the following:

The Operator registers HAProxy with --listen-port=8404 by default. If you pass --listen-port=<value>, it replaces the default 8404.
Any other flags you pass are appended to the pmm-admin add haproxy command.

Value type	Example
string	`"--custom-labels=env=prod --skip-connection-check"`

`pmm.readinessProbe.initialDelaySeconds`¶

The number of seconds to wait before performing the first readiness probe .

Value type	Example
int	`15`

`pmm.readinessProbe.timeoutSeconds`¶

The number of seconds after which the readiness probe times out.

Value type	Example
int	`15`

`pmm.readinessProbe.periodSeconds`¶

How often to perform the readiness probe . Measured in seconds.

Value type	Example
int	`30`

`pmm.readinessProbe.successThreshold`¶

The number of successful probes required to mark the container successful.

Value type	Example
int	`1`

`pmm.readinessProbe.failureThreshold`¶

The number of failed probes required to mark the container unready.

Value type	Example
int	`5`

`pmm.livenessProbe.initialDelaySeconds`¶

The number of seconds to wait before performing the first liveness probe .

Value type	Example
int	`300`

`pmm.livenessProbe.timeoutSeconds`¶

The number of seconds after which the liveness probe times out.

Value type	Example
int	`5`

`pmm.livenessProbe.periodSeconds`¶

How often to perform the liveness probe . Measured in seconds.

Value type	Example
int	`10`

`pmm.livenessProbe.successThreshold`¶

The number of successful probes required to mark the container successful.

Value type	Example
int	`1`

`pmm.livenessProbe.failureThreshold`¶

The number of failed probes required to mark the container unhealthy.

Value type	Example
int	`3`

`pmm.resources.requests.memory`¶

The Kubernetes memory requests for a PMM container.

Value type	Example
string	`150M`

`pmm.resources.requests.cpu`¶

Kubernetes CPU requests for a PMM container.

Value type	Example
string	`300m`

`pmm.resources.limits.memory`¶

Kubernetes memory limits for a PMM container.

Value type	Example
string	`256M`

`pmm.resources.limits.cpu`¶

Kubernetes CPU limits for a PMM container.

Value type	Example
string	`400m`

`pmm.serverHost`¶

Address of the PMM Server to collect data from the cluster.

Value type	Example
string	`monitoring-service`

`pmm.customClusterName`¶

A custom name to define for a cluster. PMM Server uses this name to properly parse the metrics and display them on dashboards. Using a custom name is useful for clusters deployed in different data centers - PMM Server connects them and monitors them as one deployment. Another use case is for clusters deployed with the same name in different namespaces - PMM treats each cluster separately.

Value type	Example
string	`cluster1-custom`

Log collector section¶

The logcollector section in the deploy/cr.yaml file configures the Fluent Bit and logrotate sidecars on MySQL Pods. Available starting with Operator version 1.3.0. See Persistent logging and Log rotation.

The Operator adds the sidecars only when crVersion is 1.3.0 or later and logcollector.enabled is true. resources and containerSecurityContext apply to both the logs and logrotate containers. image is required when the logcollector section is present.

`logcollector.enabled`¶

Enables or disables persistent logging.

Value type	Example
boolean	`true`

`logcollector.image`¶

Fluent Bit Docker image to use.

Value type	Example
string	`perconalab/fluentbit:5.1.4-1`

`logcollector.imagePullPolicy`¶

The policy used to update images for the logs and logrotate containers. Valid values are Always, IfNotPresent, and Never.

Value type	Example
string	`Always`

`logcollector.configuration`¶

Additional Fluent Bit rules. You cannot override the default configuration. Must be in Fluent Bit YAML format. The classic .conf format is not supported. An invalid snippet is ignored when the logs container starts. Changing this field restarts the MySQL Pods.

Value type	Example
subdoc

`logcollector.env.name`¶

Name of an environment variable for the logs container.

Value type	Example
string	`AWS_REGION`

`logcollector.env.value`¶

Value of an environment variable for the logs container.

Value type	Example
string	`us-east-1`

`logcollector.envFrom.configMapRef.name`¶

Name of a ConfigMap whose keys are loaded as environment variables for the logs container.

Value type	Example
string	`my-log-collector-config`

`logcollector.envFrom.secretRef.name`¶

Name of a Secret whose keys are loaded as environment variables for the logs container. Use this for credentials of a remote Fluent Bit output.

Value type	Example
string	`my-log-collector-secret`

`logcollector.resources.requests.memory`¶

Kubernetes memory requests for the logs and logrotate containers.

Value type	Example
string	`150M`

`logcollector.resources.requests.cpu`¶

Kubernetes CPU requests for the logs and logrotate containers.

Value type	Example
string	`300m`

`logcollector.resources.limits.memory`¶

Kubernetes memory limits for the logs and logrotate containers.

Value type	Example
string	`200M`

`logcollector.resources.limits.cpu`¶

Kubernetes CPU limits for the logs and logrotate containers.

Value type	Example
string	`350m`

`logcollector.containerSecurityContext`¶

A custom Kubernetes Security Context for the logs and logrotate containers.

Value type	Example
subdoc	privileged: false runAsUser: 1001 runAsGroup: 1001

`logcollector.livenessProbe`¶

A custom Kubernetes liveness probe for the logs container. When unset, the container has no liveness probe. The built-in Fluent Bit configuration listens on port 2020, so a tcpSocket probe on that port works without extra configuration.

Value type	Example
subdoc	`tcpSocket: { port: 2020 }`

`logcollector.livenessProbe.initialDelaySeconds`¶

Number of seconds to wait after the logs container starts before the first liveness probe.

Value type	Example
int	`30`

`logcollector.livenessProbe.periodSeconds`¶

How often to run the liveness probe, in seconds.

Value type	Example
int	`15`

`logcollector.readinessProbe`¶

A custom Kubernetes readiness probe for the logs container. When unset, the container has no readiness probe. The built-in Fluent Bit configuration listens on port 2020.

Value type	Example
subdoc	`tcpSocket: { port: 2020 }`

`logcollector.readinessProbe.initialDelaySeconds`¶

Number of seconds to wait after the logs container starts before the first readiness probe.

Value type	Example
int	`5`

`logcollector.readinessProbe.periodSeconds`¶

How often to run the readiness probe, in seconds.

Value type	Example
int	`10`

`logcollector.volumeMounts`¶

Extra volume mounts for the logs and logrotate containers. Use these with logcollector.volumes. For example, to mount a CA bundle for an S3 output.

Value type	Example
subdoc	- name: s3-ca mountPath: /etc/fluentbit/tls readOnly: true

`logcollector.volumes`¶

Extra volumes on the MySQL Pod for the log collector sidecars.

Value type	Example
subdoc	- name: s3-ca secret: secretName: my-s3-ca

`logcollector.logRotate.configuration`¶

Replaces the default logrotate configuration. You must provide the full configuration. An invalid snippet is ignored and the sidecar uses the built-in rules. See Override the default configuration.

Value type	Example
subdoc

`logcollector.logRotate.extraConfig.name`¶

Name of a ConfigMap in the same namespace that holds extra logrotate rules. Each key must end with .conf. Do not use mysql.conf as the name. That name is reserved for the Operator-managed configuration. See Add extra rules from a ConfigMap.

Value type	Example
string	`my-logrotate-extra`

`logcollector.logRotate.schedule`¶

Cron expression of five fields for the logrotate sidecar. The default is 0 0 * * * (once a day at midnight). The Operator rejects a value that is not valid cron or that contains a newline.

Value type	Example
string	`0 /6 * *`

`logcollector.logRotate.livenessProbe`¶

A custom Kubernetes liveness probe for the logrotate container. When unset, the container has no liveness probe.

Value type	Example
subdoc	`exec: { command: [/bin/true] }`

`logcollector.logRotate.livenessProbe.initialDelaySeconds`¶

Number of seconds to wait after the logrotate container starts before the first liveness probe.

Value type	Example
int	`30`

`logcollector.logRotate.livenessProbe.periodSeconds`¶

How often to run the liveness probe, in seconds.

Value type	Example
int	`15`

`logcollector.logRotate.readinessProbe`¶

A custom Kubernetes readiness probe for the logrotate container. When unset, the container has no readiness probe.

Value type	Example
subdoc	`exec: { command: [/bin/true] }`

`logcollector.logRotate.readinessProbe.initialDelaySeconds`¶

Number of seconds to wait after the logrotate container starts before the first readiness probe.

Value type	Example
int	`5`

`logcollector.logRotate.readinessProbe.periodSeconds`¶

How often to run the readiness probe, in seconds.

Value type	Example
int	`10`

Backup section¶

The backup section in the deploy/cr.yaml file contains the following configuration options for the regular Percona XtraDB Cluster backups.

`backup.enabled`¶

Enables or disables making backups.

Value type	Example
boolean	`true`

`backup.encryptionKeySecret.name`¶

The name of the Kubernetes Secret that stores the key used to encrypt backups before they are stored in object storage. When set, all backup storages use this key unless you specify a different Secret for a specific storage.

Value type	Example
string	`my-encryption-key`

`backup.encryptionKeySecret.key`¶

The key within the Secret that holds the encryption key value. Defaults to encryptionKey. This is a global setting and applies to all storages unless overridden for a specific storage.

Value type	Example
string	`encryptionKey`

`backup.pitr.enabled`¶

Enables or disables point-in-time recovery functionality.

Value type	Example
boolean	`true`

`backup.pitr.binlogServer.size`¶

Controls the number of Percona Binarylog Server Pods for binlog collection. Defaults to 1. The Custom Resource allows only 1 (larger values are rejected). To learn more, see Point-in-time recovery.

Value type	Example
int	`1`

`backup.pitr.binlogServer.image`¶

The Docker image to use to deploy Percona Binarylog Server.

Value type	Example
string	`perconalab/percona-binlog-server:0.2.1`

`backup.pitr.binlogServer.imagePullPolicy`¶

The policy the kubelet uses to pull the Percona Binarylog Server image.

Value type	Example
string	`Always`

`backup.pitr.binlogServer.imagePullSecrets.name`¶

The name of a Secret the kubelet uses to authenticate to a private image registry. You can list several Secrets as an array of objects with the name key.

Value type	Example
string	`my-secret-1`

`backup.pitr.binlogServer.serverId`¶

The unique server ID that Percona Binarylog Server uses when connecting to MySQL as a replication client for binlog collection.

Value type	Example
int	`100`

`backup.pitr.binlogServer.storage.s3.bucket`¶

The name of the bucket on AWS S3 or S3-compatible storage where binlogs are streamed.

Value type	Example
string	`S3-BACKUP-BUCKET-NAME-HERE`

`backup.pitr.binlogServer.storage.s3.credentialsSecret`¶

The Kubernetes Secret for binlog storage. It should contain AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY keys.

Value type	Example
string	`ps-cluster1-s3-credentials`

`backup.pitr.binlogServer.storage.s3.endpointUrl`¶

The URL to access the bucket on S3-compatible storage. Not needed for AWS S3.

Value type	Example
string	`https://s3.amazonaws.com`

`backup.pitr.binlogServer.storage.s3.prefix`¶

The path prefix (folder) in the bucket where binlogs are stored. You cannot change the prefix after you configured the Binlog Server.

Value type	Example
string	`PREFIX_NAME`

`backup.pitr.binlogServer.storage.s3.region`¶

The region of the bucket. Required for Amazon S3 and for S3-compatible storage.

Value type	Example
string	`us-west-2`

`backup.pitr.binlogServer.storage.s3.caBundle.name`¶

The name of the Secret that stores the CA certificate used to verify TLS communication with S3-compatible binlog storage. This setting is independent from backup.storages.STORAGE-NAME.s3.caBundle. To change the bundle, create a new Secret and set this field to its name. See Update the CA bundle.

Value type	Example
string	`minio-ca-bundle`

`backup.pitr.binlogServer.storage.s3.caBundle.key`¶

The key in the Secret that holds the CA certificate. If you omit this field, the Operator defaults to ca.crt.

Value type	Example
string	`ca.crt`

`backup.pitr.binlogServer.storage.encryption.cipher`¶

The cipher used to encrypt binlog data before Percona Binlog Server writes files to object storage. Only CTR mode is supported.

Allowed values: AES-128-CTR, AES-192-CTR, AES-256-CTR. Defaults to AES-256-CTR.

Value type	Example
string	`AES-256-CTR`

`backup.pitr.binlogServer.storage.encryption.kekId`¶

The id of the key encryption key (KEK) in the keyring Secret used to wrap the per-file data encryption key. If omitted, Percona Binlog Server uses the first key in the keyring. Changing this value affects only new binlogs. See Binlog encryption.

Value type	Example
string	`alpha`

`backup.pitr.binlogServer.keyringSecret.name`¶

The name of the Kubernetes Secret that stores the binlog keyring file. Required when storage.encryption is set. Keep this Secret configured after you disable encryption if the bucket still contains encrypted binlogs. See Binlog encryption.

Value type	Example
string	`ps-cluster1-binlog-server-keyring`

`backup.pitr.binlogServer.keyringSecret.key`¶

The key within the Secret that holds the keyring JSON. Defaults to keyring.json.

Value type	Example
string	`keyring.json`

`backup.pitr.binlogServer.connectTimeout`¶

Timeout in seconds for establishing a connection to MySQL.

Value type	Example
int	`30`

`backup.pitr.binlogServer.readTimeout`¶

The maximum time in seconds the Binlog Server waits to read data from the MySQL instance.

Value type	Example
int	`30`

`backup.pitr.binlogServer.writeTimeout`¶

The maximum time in seconds the Binlog Server waits to write data to a remote server.

Value type	Example
int	`30`

`backup.pitr.binlogServer.idleTime`¶

The maximum time in seconds the Binlog Server stays in idle mode before trying to reconnect.

Value type	Example
int	`30`

`backup.pitr.binlogServer.affinity.antiAffinityTopologyKey`¶

The Kubernetes topologyKey constraint for node anti-affinity on Percona Binarylog Server Pods.

Value type	Example
string	`kubernetes.io/hostname`

`backup.pitr.binlogServer.affinity.advanced`¶

In cases where the pods require complex tuning the advanced option turns off the topologykey effect. This setting allows the standard Kubernetes affinity constraints of any complexity to be used.

Value type	Example
subdoc

`backup.pitr.binlogServer.checkpointInterval`¶

How often the Binlog Server writes checkpoints - the snapshot of its state.

Value type	Example
string	`30s`

`backup.pitr.binlogServer.checkpointSize`¶

Defines the size threshold for checkpoints. Default is 16M.

Value type	Example
string	`16M`

`backup.pitr.binlogServer.containerSecurityContext`¶

A custom Kubernetes security context for a container for the Percona Binlog Server container.

Value type	Example
subdoc	privileged: false runAsUser: 1001 runAsGroup: 1001

`backup.pitr.binlogServer.env`¶

Custom environment variables for the Percona Binarylog Server container.

Value type	Example
subdoc	`[]`

`backup.pitr.binlogServer.envFrom`¶

The source for the bulk-import of all key-value pairs from a ConfigMap or Secret as environment variables for the Percona Binarylog Server container.

Value type	Example
subdoc	`[]`

`backup.pitr.binlogServer.initContainer.image`¶

The container image for the Binlog Server init container.

Value type	Example
string	`perconalab/percona-server-mysql-operator:main`

`backup.pitr.binlogServer.initContainer.containerSecurityContext`¶

A custom Kubernetes security context for a container for the Binlog Server init container.

Value type	Example
subdoc	privileged: false runAsUser: 1001 runAsGroup: 1001

`backup.pitr.binlogServer.initContainer.resources.requests.memory`¶

Kubernetes memory requests for the Binlog Server init container.

Value type	Example
string	`200M`

`backup.pitr.binlogServer.initContainer.resources.requests.cpu`¶

Kubernetes CPU requests for the Binlog Server init container.

Value type	Example
string	`200m`

`backup.pitr.binlogServer.initContainer.resources.limits.memory`¶

Kubernetes memory limits for the Binlog Server init container.

Value type	Example
string	`100M`

`backup.pitr.binlogServer.initContainer.resources.limits.cpu`¶

Kubernetes CPU limits for the Binlog Server init container.

Value type	Example
string	`100m`

`backup.pitr.binlogServer.initImage`¶

An alternative init image for the Percona Binarylog Server Pod. Leave empty to use the default one.

Value type	Example
string	`""`

`backup.pitr.binlogServer.logLevel`¶

Log level for the Percona Binarylog Server. Defaults to info.

Value type	Example
string	`info`

`backup.pitr.binlogServer.podSecurityContext`¶

A custom Kubernetes Pod security context for Percona Binarylog Server Pods.

Value type	Example
subdoc	fsGroup: 1001 supplementalGroups: - 1001

`backup.pitr.binlogServer.resources.requests.memory`¶

Kubernetes memory requests for the Percona Binarylog Server container.

Value type	Example
string	`200M`

`backup.pitr.binlogServer.resources.requests.cpu`¶

Kubernetes CPU requests for the Percona Binarylog Server container.

Value type	Example
string	`200m`

`backup.pitr.binlogServer.resources.limits.memory`¶

Kubernetes memory limits for the Percona Binarylog Server container.

Value type	Example
string	`100M`

`backup.pitr.binlogServer.resources.limits.cpu`¶

Kubernetes CPU limits for the Percona Binarylog Server container.

Value type	Example
string	`100m`

`backup.pitr.binlogServer.rewriteFileSize`¶

Specifies the maximum binlog file size for rewrite. Defaults to 128M. Set to an empty string to use the default.

Value type	Example
string	`128M`

`backup.pitr.binlogServer.sslMode`¶

The TLS mode for the Binlog Server connection to MySQL. Defaults to verify_identity.

Value type	Example
string	`verify_identity`

`backup.sourcePod`¶

Specifies the MySQL instance Pod to take a backup from. When defined, takes precedence, regardless of the cluster type (async or group-replication) and topology. Applies both to scheduled and on-demand backups.

Asynchronous replication clusters that consist of more than one Pod and have the Orchestrator disabled must have the sourcePod defined for the Operator to make backups. Otherwise, the Operator fails to start a backup and reports an error.

Value type	Example
string	`ps-cluster1-mysql-0`

`backup.image`¶

The Percona XtraBackup Docker image to use for the backup.

Value type	Example
string	`percona/percona-server-mysql-operator:1.3.0-backup`

`backup.imagePullPolicy`¶

The policy used to update images .

`backup.imagePullSecrets.name`¶

The Kubernetes ImagePullSecret .

Value type	Example
string	`my-secret-1`

`backup.initContainer.image`¶

An alternative init image for Percona XtraBackup Pods.

Value type	Example
string	`perconalab/percona-server-mysql-operator:1.3.0`

`backup.initContainer.containerSecurityContext`¶

A custom Kubernetes Security Context for a Container for the image used for Percona XtraBackup Pods installation.

Value type	Example
subdoc	privileged: false runAsUser: 1001 runAsGroup: 1001

`backup.initContainer.resources.requests.memory`¶

The Kubernetes memory requests for an image used for Percona XtraBackup Pods installation.

Value type	Example
string	`100M`

`backup.initContainer.resources.requests.cpu`¶

Kubernetes CPU requests for an image used for Percona XtraBackup Pods installation.

Value type	Example
string	`100m`

`backup.initContainer.resources.limits.memory`¶

Kubernetes memory limits for an image used for Percona XtraBackup Pods installation.

Value type	Example
string	`200M`

`backup.initContainer.resources.limits.cpu`¶

Kubernetes CPU limits for an image used for Percona XtraBackup Pods installation.

Value type	Example
string	`200m`

`backup.containerSecurityContext`¶

A custom Kubernetes Security Context for the xtrabackup container to be used instead of the default one.

Value type	Example
subdoc	`privileged: true`

`backup.backoffLimit`¶

The number of retries to make a backup (by default, 6 retries are made).

Value type	Example
int	`6`

`backup.startingDeadlineSeconds`¶

The maximum time, in seconds, a backup can wait to start before the Operator marks it as Error. The backup can wait if the cluster isn’t ready yet or another backup for this cluster is still running. The default is 0, which means the Operator does not enforce a start deadline. Applies to all backups for this cluster. Overridable per backup with PerconaServerMySQLBackup.spec.startingDeadlineSeconds.

Value type	Example
int	`900`

`backup.suspendedDeadlineSeconds`¶

The maximum time, in seconds, a suspended backup can wait to resume before the Operator marks it as failed. The default is 0, which means the Operator does not enforce a resume deadline.

Applies to all backups for this cluster. Overridable per backup with PerconaServerMySQLBackup.spec.suspendedDeadlineSeconds.

Value type	Example
int	`900`

`backup.storages.STORAGE-NAME.type`¶

The cloud storage type used for backups. The following types are supported: s3, gcs and azure.

Value type	Example
string	`s3`

`backup.storages.STORAGE-NAME.verifyTLS`¶

Enable or disable verification of the storage server TLS certificate. Disabling it may be useful e.g. to skip TLS verification for private S3-compatible storage with a self-issued certificate.

Starting with version 1.3.0, the Operator supports TLS verification with your organization’s CA. Keep this option enabled and use the CA your company already trusts. See Configure TLS verification with custom certificates for S3 storage.

Value type	Example
boolean	`true`

`backup.storages.STORAGE-NAME.encryptionKeySecret.name`¶

The name of the Kubernetes Secret that contains the key used to encrypt backups for this storage location. When set, it takes precedence over backup.encryptionKeySecret.name.

Value type	Example
string	`my-encryption-key`

`backup.storages.STORAGE-NAME.encryptionKeySecret.key`¶

The key within the Secret that holds the encryption key value. Defaults to encryptionKey. It takes precedence over the backup.encryptionKeySecret.key setting.

Value type	Example
string	`encryptionKey`

`backup.storages.STORAGE-NAME.nodeSelector`¶

Kubernetes nodeSelector .

Value type	Example
label	`disktype: ssd`

`backup.storages.STORAGE-NAME.resources.requests.memory`¶

The Kubernetes memory requests for a Percona XtraBackup container.

Value type	Example
string	`1G`

`backup.storages.STORAGE-NAME.resources.requests.cpu`¶

Kubernetes CPU requests for a Percona XtraBackup container.

Value type	Example
string	`600m`

`backup.storages.STORAGE-NAME.affinity.nodeAffinity`¶

The Operator node affinity constraint.

Value type	Example
subdoc

`backup.storages.STORAGE-NAME.topologySpreadConstraints.labelSelector.matchLabels`¶

The Label selector for the Kubernetes Pod Topology Spread Constraints .

Value type	Example
label	`app.kubernetes.io/name: percona-server`

`backup.storages.STORAGE-NAME.topologySpreadConstraints.maxSkew`¶

The degree to which Pods may be unevenly distributed under the Kubernetes Pod Topology Spread Constraints .

Value type	Example
int	1

`backup.storages.STORAGE-NAME.topologySpreadConstraints.topologyKey`¶

The key of node labels for the Kubernetes Pod Topology Spread Constraints .

Value type	Example
string	`kubernetes.io/hostname`

`backup.storages.STORAGE-NAME.topologySpreadConstraints.whenUnsatisfiable`¶

What to do with a Pod if it doesn’t satisfy the Kubernetes Pod Topology Spread Constraints .

Value type	Example
string	`DoNotSchedule`

`backup.storages.STORAGE-NAME.tolerations`¶

Kubernetes Pod tolerations .

Value type	Example
subdoc

`backup.storages.STORAGE-NAME.schedulerName`¶

The Kubernetes Scheduler .

Value type	Example
string	`mycustom-scheduler`

`backup.storages.STORAGE-NAME.priorityClassName`¶

The Kubernetes Pod priority class .

Value type	Example
string	`high-priority`

`backup.storages.STORAGE-NAME.containerSecurityContext`¶

A custom Kubernetes Security Context for a Container to be used instead of the default one.

Value type	Example
subdoc	`privileged: true`

`backup.storages.STORAGE-NAME.podSecurityContext`¶

A custom Kubernetes Security Context for a Pod to be used instead of the default one.

Value type	Example
subdoc	fsGroup: 1001 supplementalGroups: [1001, 1002, 1003]

`backup.storages.STORAGE-NAME.runtimeClassName`¶

Specifies the name of the RuntimeClass resource used to define and select the container runtime configuration for backup and restore jobs associated with the specific storage.

Value type	Example
string	`image-rc`

`backup.storages.STORAGE-NAME.containerOptions.env`¶

The environment variables set as key-value pairs for the backup container.

Value type	Example
subdoc	- name: CUSTOM_ENV value: “some-value”

`backup.storages.STORAGE-NAME.containerOptions.args.xtrabackup`¶

Custom command line options for the xtrabackup Percona XtraBackup tool .

Value type	Example
subdoc	- “–someflag=abc”

`backup.storages.STORAGE-NAME.containerOptions.args.xbcloud`¶

Custom command line options for the xbcloud Percona XtraBackup tool .

Value type	Example
subdoc	- “–someflag=abc”

`backup.storages.STORAGE-NAME.containerOptions.args.xbstream`¶

Custom command line options for the xbstream Percona XtraBackup tool

Value type	Example
subdoc	- “–someflag=abc”

`backup.storages.STORAGE-NAME.annotations`¶

The Kubernetes annotations .

Value type	Example
label	`testName: scheduled-backup`

`backup.storages.STORAGE-NAME.labels`¶

Labels are key-value pairs attached to objects .

Value type	Example
label	`backupWorker: 'True'`

`backup.storages.STORAGE-NAME.s3.bucket`¶

The Amazon S3 bucket name for backups.

Value type	Example
string

`backup.storages.STORAGE-NAME.s3.region`¶

The AWS region to use. Please note this option is mandatory for Amazon and all S3-compatible storages.

Value type	Example
string	`us-west-2`

`backup.storages.STORAGE-NAME.s3.prefix`¶

The path (sub-folder) to the backups inside the bucket .

Value type	Example
string	`""`

`backup.storages.STORAGE-NAME.s3.credentialsSecret`¶

The Kubernetes secret for backups. It should contain AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY keys.

Value type	Example
string	`my-cluster-name-backup-s3`

`backup.storages.STORAGE-NAME.s3.endpointUrl`¶

The endpoint URL of the S3-compatible storage to be used (not needed for the original Amazon S3 cloud)

Value type	Example
string

`backup.storages.STORAGE-NAME.s3.caBundle.name`¶

The name of the Secret that stores the CA certificate used to verify TLS communication with S3-compatible storage. To change the bundle, create a new Secret and set this field to its name. See Update the CA bundle.

Value type	Example
string	`minio-ca-bundle`

`backup.storages.STORAGE-NAME.s3.caBundle.key`¶

The key in the Secret that holds the CA certificate. If you omit this field, the Operator uses ca.crt. The Operator mounts only this key from the Secret.

Value type	Example
string	`ca.crt`

`backup.storages.STORAGE-NAME.azure.container`¶

The name of the Microsoft Azure blob container for backups.

Value type	Example
string	`""`

`backup.storages.STORAGE-NAME.azure.prefix`¶

The path (sub-folder) to the backups inside the container .

Value type	Example
string	`""`

`backup.storages.STORAGE-NAME.azure.endpointUrl`¶

The endpoint URL of the S3-compatible storage to be used (not needed for the original Amazon S3 cloud)

Value type	Example
string	`https://accountName.blob.core.windows.net`

`backup.storages.STORAGE-NAME.azure.credentialsSecret`¶

The Kubernetes secret for backups. It should contain the AZURE_STORAGE_ACCOUNT_NAME and the AZURE_STORAGE_ACCOUNT_KEY keys.

Value type	Example
string	`my-cluster-name-backup-azure`

`backup.storages.STORAGE-NAME.gcs.bucket`¶

The name of the storage bucket.

Value type	Example
string	`bucket-name`

`backup.storages.STORAGE-NAME.gcs.prefix`¶

The path to the data directory in the bucket. If undefined, backups are stored in the bucket’s root directory.

Value type	Example
string	`prefix-name`

`backup.storages.STORAGE-NAME.gcs.endpointUrl`¶

The URL to access the data in Google Cloud Storage.

Value type	Example
string	`"https://storage.googleapis.com"`

`backup.storages.STORAGE-NAME.gcs.credentialsSecret`¶

The Kubernetes secret for backups. It should contain the ACCESS_KEY_ID and the SECRET_ACCESS_KEY keys.

Value type	Example
string	`ps-cluster1-gcp-credentials`

`backup.schedule.name`¶

Name of the scheduled backup.

Value type	Example
string	`sat-night-backup`

`backup.schedule.schedule`¶

Scheduled time of the backup, specified in the crontab format .

Value type	Example
string	`0 0 \* \* 6`

`backup.schedule.keep`¶

The amount of most recent backups to store. Older backups are automatically deleted. Set keep to zero or completely remove it to disable automatic deletion of backups. Note that keep is ignored for incremental backups.

Value type	Example
int	`3`

`backup.schedule.storageName`¶

The name of the storage for the backups configured in the storages subsection.

Value type	Example
string	`s3-us-west`

`backup.schedule.type`¶

The backup type. Supported values: full, incremental. When undefined, defaults to full.

Value type	Example
string	`full`

Percona Toolkit section¶

The toolkit section in the deploy/cr.yaml file contains configuration options for Percona Toolkit .

`toolkit.image`¶

Percona Toolkit client Docker image to use.

Value type	Example
string	`percona/pmm-client:3.10.0`

`toolkit.imagePullPolicy`¶

The policy used to update images .

Value type	Example
string	`Always`

`toolkit.resources.requests.memory`¶

The Kubernetes memory requests for a Percona Toolkit container.

Value type	Example
string	`150M`

`toolkit.resources.requests.cpu`¶

Kubernetes CPU requests for a Percona Toolkit container.

Value type	Example
string	`100m`

`toolkit.resources.limits.memory`¶

Kubernetes memory limits for a Percona Toolkit container.

Value type	Example
string	`256M`

`toolkit.resources.limits.cpu`¶

Kubernetes CPU limits for a Percona Toolkit container

Value type	Example
string	`400m`

Last update: October 29, 2026
Created: September 19, 2022

Custom Resource options reference¶

PerconaServerMySQL Custom Resource options¶

Toplevel spec elements¶

crVersion¶

metadata.annotations¶

metadata.labels¶

pause¶

enableVolumeExpansion¶

Storage scaling section¶

storageScaling.enableVolumeScaling¶

storageScaling.enableExternalAutoscaling¶

storageScaling.autoscaling.enabled¶

storageScaling.autoscaling.triggerThresholdPercent¶

storageScaling.autoscaling.growthStep¶

storageScaling.autoscaling.maxSize¶

initContainer.Image¶

initContainer.containerSecurityContext¶

initContainer.resources.requests.memory¶

initContainer.resources.requests.cpu¶

initContainer.resources.limits.memory¶

initContainer.resources.limits.cpu¶

secretsName¶

clusterServiceDNSSuffix¶

sslSecretName¶

ignoreAnnotations¶

ignoreLabels¶

updateStrategy¶

pause¶

allowUnsafeConfigurations¶

Unsafe flags section¶

unsafeFlags.asyncFailoverWithPossibleDataLoss¶

unsafeFlags.backupNonReadyCluster¶

unsafeFlags.mysqlSize¶

unsafeFlags.proxy¶

unsafeFlags.proxySize¶

unsafeFlags.orchestrator¶

unsafeFlags.orchestratorSize¶

Extended cert-manager configuration section¶

tls.SANs¶

tls.issuerConf.name¶

tls.issuerConf.kind¶

tls.issuerConf.group¶

Upgrade options section¶

upgradeOptions.versionServiceEndpoint¶

upgradeOptions.apply¶

Percona Server for MySQL section¶

mysql.clusterType¶

mysql.bootstrap.mode¶

mysql.autoRecovery¶

mysql.vaultSecretName¶

mysql.size¶

mysql.image¶

mysql.imagePullPolicy¶

mysql.runtimeClassName¶

mysql.schedulerName¶

mysql.priorityClassName¶

mysql.nodeSelector¶

mysql.serviceAccountName¶

mysql.tolerations¶

mysql.imagePullSecrets.name¶

mysql.initContainer.image¶

mysql.initContainer.containerSecurityContext¶

mysql.initContainer.resources.requests.memory¶

mysql.initContainer.resources.requests.cpu¶

mysql.initContainer.resources.limits.memory¶

mysql.initContainer.resources.limits.cpu¶

mysql.podDisruptionBudget.maxUnavailable¶

mysql.podDisruptionBudget.minAvailable¶

mysql.resources.requests.memory¶

mysql.resources.requests.cpu¶

mysql.resources.limits.memory¶

mysql.resources.limits.cpu¶

mysql.startupProbe.initialDelaySeconds¶

mysql.startupProbe.timeoutSeconds¶

mysql.startupProbe.periodSeconds¶

mysql.startupProbe.successThreshold¶

mysql.startupProbe.failureThreshold¶

mysql.readinessProbe.initialDelaySeconds¶

mysql.readinessProbe.timeoutSeconds¶

mysql.readinessProbe.periodSeconds¶

Toplevel `spec` elements¶

`crVersion`¶

`metadata.annotations`¶

`metadata.labels`¶

`pause`¶

`enableVolumeExpansion`¶

`storageScaling.enableVolumeScaling`¶

`storageScaling.enableExternalAutoscaling`¶

`storageScaling.autoscaling.enabled`¶

`storageScaling.autoscaling.triggerThresholdPercent`¶

`storageScaling.autoscaling.growthStep`¶

`storageScaling.autoscaling.maxSize`¶

`initContainer.Image`¶

`initContainer.containerSecurityContext`¶

`initContainer.resources.requests.memory`¶

`initContainer.resources.requests.cpu`¶

`initContainer.resources.limits.memory`¶

`initContainer.resources.limits.cpu`¶

`secretsName`¶

`clusterServiceDNSSuffix`¶

`sslSecretName`¶

`ignoreAnnotations`¶

`ignoreLabels`¶

`updateStrategy`¶

`pause`¶

`allowUnsafeConfigurations`¶

`unsafeFlags.asyncFailoverWithPossibleDataLoss`¶

`unsafeFlags.backupNonReadyCluster`¶

`unsafeFlags.mysqlSize`¶

`unsafeFlags.proxy`¶

`unsafeFlags.proxySize`¶

`unsafeFlags.orchestrator`¶

`unsafeFlags.orchestratorSize`¶

`tls.SANs`¶

`tls.issuerConf.name`¶

`tls.issuerConf.kind`¶

`tls.issuerConf.group`¶

`upgradeOptions.versionServiceEndpoint`¶

`upgradeOptions.apply`¶

`mysql.clusterType`¶

`mysql.bootstrap.mode`¶

`mysql.autoRecovery`¶

`mysql.vaultSecretName`¶

`mysql.size`¶

`mysql.image`¶

`mysql.imagePullPolicy`¶

`mysql.runtimeClassName`¶

`mysql.schedulerName`¶

`mysql.priorityClassName`¶

`mysql.nodeSelector`¶

`mysql.serviceAccountName`¶

`mysql.tolerations`¶

`mysql.imagePullSecrets.name`¶

`mysql.initContainer.image`¶

`mysql.initContainer.containerSecurityContext`¶

`mysql.initContainer.resources.requests.memory`¶

`mysql.initContainer.resources.requests.cpu`¶

`mysql.initContainer.resources.limits.memory`¶

`mysql.initContainer.resources.limits.cpu`¶

`mysql.podDisruptionBudget.maxUnavailable`¶

`mysql.podDisruptionBudget.minAvailable`¶

`mysql.resources.requests.memory`¶

`mysql.resources.requests.cpu`¶

`mysql.resources.limits.memory`¶

`mysql.resources.limits.cpu`¶

`mysql.startupProbe.initialDelaySeconds`¶

`mysql.startupProbe.timeoutSeconds`¶

`mysql.startupProbe.periodSeconds`¶

`mysql.startupProbe.successThreshold`¶

`mysql.startupProbe.failureThreshold`¶

`mysql.readinessProbe.initialDelaySeconds`¶

`mysql.readinessProbe.timeoutSeconds`¶

`mysql.readinessProbe.periodSeconds`¶

`mysql.readinessProbe.successThreshold`¶

`mysql.readinessProbe.failureThreshold`¶

`mysql.livenessProbe.initialDelaySeconds`¶

`mysql.livenessProbe.timeoutSeconds`¶

`mysql.livenessProbe.periodSeconds`¶

`mysql.livenessProbe.successThreshold`¶

`mysql.livenessProbe.failureThreshold`¶