What it does
Percona Monitoring and Management (PMM) is a self-hosted monitoring and query-analysis system for database environments. A central PMM Server presents dashboards and stores telemetry, while PMM Clients and exporters collect database and host information. Teams can inspect metrics for MySQL, PostgreSQL, MongoDB, and Valkey or Redis deployments, then use Query Analytics to narrow a performance problem from a server or cluster down to individual database statements. PMM is aimed at database operators who want to run the monitoring stack on infrastructure they control rather than send its monitoring data to a separate hosted observability account.
Follow a database signal to a query
PMM separates collection from analysis. First deploy the PMM Server, commonly as a Docker container for an evaluation, or choose an installation path that fits an existing Kubernetes, Podman, or virtual-machine environment. Then install a PMM Client near the database host and register the database service with credentials that have the permissions required by its collector. Once data is flowing, the web interface provides host and database dashboards as well as Query Analytics.
That split is useful when a database appears slow but the source is unclear: an operator can compare host resource charts with database activity, then inspect query-level measurements to decide which statements deserve investigation. The documentation describes this workflow; no PMM instance or database was configured for this listing.
PMM is software to operate, not an external alerting service that removes infrastructure work. The team must size and maintain the server, retain its telemetry, protect the web interface, and allow the client to reach both the server and monitored database. Percona describes the project as having no paid feature edition, but costs for the compute and storage chosen to run it remain deployment-dependent.
Strengths
- Connects host and database measurements with query-level analysis, which helps narrow a database performance investigation beyond a single server metric.
- Supports self-managed deployment choices including Docker, Podman, virtual appliances, and Kubernetes or OpenShift through Helm.
- The upstream project supports several database families, including MySQL, PostgreSQL, MongoDB, and Valkey or Redis.
Watch out for
- The operator must deploy and maintain a central server and clients, provide persistent storage, and configure access between PMM and the monitored databases.
Best for
- Database administrators tracing a slowdown from host or cluster symptoms to query behavior
- Teams monitoring MySQL, PostgreSQL, MongoDB, or Valkey and Redis systems on infrastructure they manage
Sources
- Monitoring - Percona (Oct 26, 2026)
- Quickstart guide - Percona Monitoring and Management (Oct 26, 2026)
- percona/pmm - Percona Monitoring and Management (Oct 26, 2026)
- PMM Server GNU Affero General Public License v3 (Oct 26, 2026)