Configure environment variables for PMM Server on AWS

Set up environment variables in the systemd environment file to customize performance, storage, features, and other settings without modifying the container directly.

Your PMM Server on AWS runs as a systemd user service that launches a Podman container, with environment variables configured through a dedicated environment file.

Configure environment variables

To set environment variables for PMM Server on AWS:

1. Connect to your PMM Server instance via SSH using the admin user:

   ssh -i your-key.pem admin@<pmm-server-ip>
   

2. Open /home/admin/.config/systemd/user/pmm-server.env and edit the environment variables file. This file is automatically loaded by the systemd service:

   nano ~/.config/systemd/user/pmm-server.env
   

3. Add your desired environment variables in KEY=VALUE format:

   PMM_DEBUG=true
   PMM_ENABLE_ACCESS_CONTROL=true
   

4. Restart the PMM Server service to apply the changes:

   systemctl --user restart pmm-server
   

5. Verify the service is running with the new configuration:

   systemctl --user status pmm-server
   

Available variables

Unlike Docker deployments that use -e flags, AWS AMI instances configure PMM Server using the systemd environment file instructions above. However, PMM uses the same list of environment variables across all deployment methods.

Common examples for AWS AMI

These variables are particularly useful for AWS AMI deployments:

Variable	Default	Description
PMM_DATA_RETENTION	30d	Duration to retain metrics data
PMM_ENABLE_UPDATES	true	Allow version checks and updates
PMM_ENABLE_ACCESS_CONTROL	true	Enable label-based access control (LBAC)
PMM_PUBLIC_ADDRESS	Auto-detected	External DNS/IP for PMM Server
PMM_DEBUG	true	Enable verbose logging
PMM_METRICS_RESOLUTION	1s	Base metrics collection interval

VictoriaMetrics performance tuning

PMM Server uses VictoriaMetrics as its metrics storage engine. For high-volume environments or extended retention periods, you may need to tune VictoriaMetrics settings to optimize performance and resource usage. To do this:

1. Add the following variables to your pmm-server.env file:

   # Configure disk space limit per client during network outages
   VMAGENT_remoteWrite_maxDiskUsagePerURL=52428800
   
   # Configure temporary data storage path
   VMAGENT_remoteWrite_tmpDataPath=/tmp/custom-vmagent
   
   # Configure logging verbosity level
   VMAGENT_loggerLevel=DEBUG
   
   # Configure maximum scrape size per target
   VMAGENT_promscrape_maxScrapeSize=128MiB
   

2. Restart the PMM Server service after changing environment variables for VictoriaMetrics and monitor disk space when extending retention periods.

When to configure these settings:

• Large deployments (>100 monitored nodes): increase query timeout and adjust memory limits
• Extended retention (>30 days): Set VM_retentionPeriod independently from PMM_DATA_RETENTION for finer control
• Memory-constrained environments: reduce VM_memory.allowedPercent to prevent OOM issues
• High write volume: add merge speed optimization for better ingestion performance