Monitoring checklist | Postgres.FM 007 | #PostgreSQL #Postgres podcast

[Please subscribe, like, and share the materials in your social networks and groups!] Nikolay takes us through a checklist of important things to monitor, while Michael tries to keep up.

Monitoring checklist (dashboard #1):

1. TPS and (optional but also desired) QPS
2. Latency (query duration) — at least average. Better: histogram, percentiles
3. Connections (sessions) — stacked graph of session counts by state (first of all: active and idle-in-transaction; also interesting: idle, others) and how far the sum is from max_connection (+pool size for PgBouncer).
4. Longest transactions (max transaction age or top-n transactions by age), excluding autovacuum activity
5. Commits vs rollbacks — how many transactions are rolled back
6. Transactions left till transaction ID wraparound
7. Replication lags / bytes in replication slot / unused replication slots
8. Count of WALs waiting to be archived (archiving lag)
9. WAL generation rates
10. Locks and deadlocks
11. Basic query analysis graph (top-n by total_time or by mean_time?)
12. Basic wait event analysis (a.k.a. “active session analysis” or “performance insights”)

And links to a few things we mentioned:

------------------------

~~~
Postgres FM is brought to you by:
- Michael Christofides, founder of pgMustard

~~~

[magfal]

I target 100 ms including network round trips for user interactive queries.
This makes pgMustard a great tool for my use case.

[shahkeyu]

Thanks for this informative topic and session.
Any recommended postgres exporter for Prometheus? I already tried postgres_exporter and pg_scv