Re: Suggestion to Monitoring Tool

On Wed, May 27, 2020 at 12:46 PM postgann2020 s <postgann2020(at)gmail(dot)com>
wrote:

> Hi Team,
>
> Thanks for your support.
>
> Environment Details:
> OS: RHEL 7.2
> Postgres: 9.5.15
> Master-Slave with Streaming replication
>
> We are planning to implement the monitoring tool for our environment.
>
> Could someone please suggest the Monitoring Tool based on your experience.
>
> We are looking to cover the below areas.
>
> 1. Monitoring metrics and alerting.
> 2. Monitoring events and alerting.
> 3. Consolidate all the PROD DB logs and provide insights on log data.
> 4. logging explain plan and insights on explain plans. (Something like
> store explain plan and compare plans and send alerts on deviations)
> 5. Logging audit data and insights from audit data.
>
>
> Thanks & Regards,
> Postgann.
>

For items 1 & 2, we've been working on a monitoring solution for PostgreSQL
using Prometheus & Grafana called pgMonitor

https://github.com/CrunchyData/pgmonitor

We went this approach so it's not just confined to monitoring only
PostgreSQL. You can monitor any other part of your infrastructure as well
with the same programs.

For log monitoring (item 3), I would recommend pgBadger.
https://github.com/darold/pgbadger

To get explain plans to be part of your logs (item 4), you can look at the
auto_explain contrib module -
https://www.postgresql.org/docs/current/auto-explain.html. I'm not aware of
anything that would alert you on deviation from plans, and that would be
quite tricky to implement. You can monitor for long running queries with
pgMonitor and pgBadger can provide you with further insights on which
queries are running long.

For auditing, I would suggest the pg_audit extension -
https://github.com/pgaudit/pgaudit

So, there is not a single monitoring solution that provides all the items
you're looking for, but there are solutions to each item available.

--
Keith Fiske
Senior Database Engineer
Crunchy Data - http://crunchydata.com