Managing a fleet of PostgreSQL instances often means juggling multiple monitoring tools, each with its own agent and setup. Dasha aims to simplify this with a read-only connection to your clusters, requiring no installation on database hosts—no agents, and no extra extensions beyond those you likely already have (like pg_stat_statements). It reads every host in a cluster to provide a unified view, as statistics are per-instance and not replicated; for example, an index that looks unused on the primary might actually be serving reads on a standby.
Key features of the initial release:
- Health Score: A composite 0–100 score per instance, calculated across eight categories. It comes with prioritized recommendations and drill-downs per database. Optionally, you can compute this from Prometheus or VictoriaMetrics time series, rather than point-in-time SQL, giving you trend and seasonal baselines.
- Query Analysis: Identify top queries by execution time or WAL generation, see running and blocked queries, and work with
pg_stat_statementssnapshots. These snapshots can be saved to a storage database, shared via a URL, and diffed against other snapshots or live data. - Automatic Snapshots: A separate daemon captures query and lock-contention snapshots automatically, helping you track performance over time and spot regressions early.
One Dasha instance can serve many clusters and all their replicas, making it ideal for organizations with a large PostgreSQL footprint. It connects using a read-only role, ensuring your production data remains safe from accidental writes.
For developers, Dasha addresses a common pain point: the need for a lightweight, centralized monitoring solution that works across a fleet without deploying per-host agents. Its ability to reason over the combined picture of primary and standby instances gives you insights that are often missed by per-instance tools. The use of Prometheus/VictoriaMetrics integration for trend analysis is particularly valuable for capacity planning and anomaly detection. Being open source, you can also extend or adapt it to your specific needs.
If you're managing multiple PostgreSQL clusters, Dasha is worth exploring. It's a fresh tool that complements existing PostgreSQL observability features, making it easier to keep your databases healthy and performant.
Source: https://www.postgresql.org/about/news/dasha-performance-dashboard-3360/