Logging in Percona Link for MongoDB

Percona Link for MongoDB (PLM) provides a flexible logging system to help you monitor its operations, diagnose issues, and integrate with log management systems. You can control the log verbosity, format, and appearance using command-line flags.

Configuration

You configure logging when the plm process starts. The following flags are available:

Flag	Description	Default
--log-level	Sets the verbosity of the logs.	info
--log-json	Outputs logs in a structured JSON format.	false
--no-color	Disables colorized output for text-based logs.	false

Log level

The --log-level flag controls the minimum level of messages that will be recorded. The supported levels are, in order of severity:

• trace: Highly detailed diagnostic information.
• debug: Detailed information useful for debugging.
• info: General information about the application’s state and progress.
• warn: Potentially harmful situations or unexpected events.
• error: Errors that prevent a specific operation from completing but do not stop the application.
• fatal: Severe errors that cause the application to terminate.

Example:

To see detailed debugging messages, start plm with:

plm --source <source-uri> --target <target-uri> --log-level=debug

Log format

PLM can output logs in two formats: human-readable text (default) and structured JSON.

Text format (default)

By default, logs are printed to the console in a color-coded, human-readable format. This is ideal for interactive use and manual inspection.

Sample output

2024-10-26 14:30:01.000 INF s=http Starting HTTP server at http://localhost:2242
2024-10-26 14:30:05.123 DBG s=repl:watch op=insert ns=test.coll1 op_ts=1729953005,1

You can disable the colorization with the --no-color flag. This is useful when redirecting log output to a file.

plm --source <source-uri> --target <target-uri> --no-color > plm.log

JSON format

For automated processing and integration with log aggregation tools (like the ELK stack or Splunk), you can use the --log-json flag. This will output each log entry as a single line of JSON.

Sample output

{"level":"info","s":"http","time":"2024-10-01 14:30:01.000","message":"Starting HTTP server at http://localhost:2242"}
{"level":"debug","s":"repl:watch","op":"insert","ns":"test.coll1","op_ts":[1729953005,1],"time":"2024-10-26 14:30:05.123"}

JSON field reference

When --log-json is enabled, the following fields may appear in the log entries:

Field	Type	Description
level	string	The severity of the log entry (e.g., info, debug).
s	string	The scope or component where the log originated (e.g., http, clone, repl).
ns	string	The MongoDB namespace (database.collection) related to the event.
elapsed_secs	float	The time taken for an operation to complete, in seconds.
time	string	The timestamp of the log event in YYYY-MM-DD HH:MM:SS.ms format.
message	string	The main log message.
error	string	The error message, if an error occurred.
op	string	The type of operation (e.g., insert, createIndexes).
op_ts	array	The MongoDB operation timestamp as [timestamp, increment].
count	integer	A count of items, such as documents in a batch.
size_bytes	integer	The size of data in bytes.

---

Last update: December 2, 2025
Created: December 2, 2025