Обсуждение: [PERFORM] Monitoring tool for Postgres Database

+1 for Datadog.  It is highly configurable, but out of the box the postgres integration collects a good amount of useful stuff. Tech support is also good.

http://docs.datadoghq.com/integrations/postgresql/

Grant Evans

Enova Inc.

gevans@enova.com

On Thu, May 25, 2017 at 3:48 PM, Ravi Tammineni <rtammineni@partner.aligntech.com> wrote:

Hi,

 

What is the best monitoring tool for Postgres database? Something like Oracle Enterprise Manager.

 

Specifically I am interested in tools to help:

 

Alert DBAs to problems with both configuration and performance issues

Deadlocks, Long running queries etc.,

Monitoring of overall system performance

General performance tuning

Storage/Disk latencies

 

 

Thanks

ravi

We use Datadog.  Their PostgreSQL plugin covers most of the most relevant stats.  It is easy to configure and not very expensive at all.  They have an easy GUI based configuration for monitors and alerts, and you can link it with something like Victorops and Slack for additional pager escalation policies.  We have all of our batch processing tied into Datadog as well, so we can get a picture of events, systems, and database internals all in one dashboard.

Hi Ravi,

We, at Dalibo, are contributing to the postgres community mostly through open source administration tools development.
Among them :
- pgBadger, from Gilles Darold, extracts a lot of data from the postgres logs and restitutes them via a web interface (https://github.com/dalibo/pgbadger and http://dalibo.github.io/pgbadger/),
- OPM - Open PostgreSQL Monitoring - monitors the activity of instances and sends alerts if needed (http://opm.io/, https://github.com/OPMDG and http://opm.readthedocs.io/index.html). It uses the check_pgactivity agent,
- PoWA - PostgreSQL Workload Analyzer - captures and stores the SQL activity of instances (using the pg_stat_statements extension) and reports it through a web interface. (http://powa.readthedocs.io/en/latest/ and https://github.com/dalibo/powa). Several plugins help the DBA in understanding and improving SQL statements performance:
   - pg_qualstats evaluates the selectivity of predicates or where clause encountered in SQL statements,
   - pg_stat_kcache captures additional statistics from the OS, like CPU and physical I/Os,
   - HypoPG allows to create hypothetical indexes and shows the access plan that the postgres optimizer would choose if these indexes would exist.
- PgCluu, from Gilles Darold, performs a full audit of a PostgreSQL Cluster performances. A collector grabs statistics on the PostgreSQL cluster using psql and sar, a grapher generates all HTML and charts output. (http://pgcluu.darold.net/ and https://github.com/darold/pgcluu)

FYI, we are also working on a new project named TemBoard (http://temboard.io/ and https://github.com/dalibo/temboard). It is not yet production ready. But it has been presented at the latest postgres conference in Russia (https://pgconf.ru/en/2017/93881).

Philippe Beaudoin.

We've found pghero to be a good first line of defence. It doesn't have alerting yet, but it's great for a quick high level healthcheck. 

Also +1 for Datadog. Extremely flexible and elegant UI + powerful alerting capabilities.

On Fri, May 26, 2017 at 10:32 AM, Sunkara, Amrutha <amrutha@nytimes.com> wrote:

We have been using Nagios to monitor the system level stats. The database level stats that we gather are custom scripts that we have nagios poll to get the database health. You could use pg badger to generate reports against your database logs as well. Pg_badger reports are your bffs for performance related specs.. very close to AWR reports that oracle provides.

Sotrage/Disk latencies -- we have oracle's os watcher we running regularly on these hosts to generate iostats as well. 

Thanks.

-Amrutha.

On Thu, May 25, 2017 at 3:48 PM, Ravi Tammineni <rtammineni@partner.aligntech.com> wrote:

Hi,

 

What is the best monitoring tool for Postgres database? Something like Oracle Enterprise Manager.

 

Specifically I am interested in tools to help:

 

Alert DBAs to problems with both configuration and performance issues

Deadlocks, Long running queries etc.,

Monitoring of overall system performance

General performance tuning

Storage/Disk latencies

 

 

Thanks

ravi