Logging, monitoring, and auditing.

[jenstroeger]

Running a full-stack application requires to

• logging: collect operational runtime data (Splunk blog, Wikipedia);¹
• monitoring: observe the behavior and metrics of the runtime stack (see issue Observability of containers #8, pgExporter, PostgREST metrics, Dramatiq metrics which should cover the message queue metrics); and
• auditing: collect runtime data regarding usage events (pgAudit, Python events, Splunk blog, Wikipedia).

See also the The Subtle Differences Between Logs, Metrics, and Audits blog.

As for processing and searching through all that textual data, we then have Postgres’s text search and trigram matching built in.

Footnotes

1. There’s a bunch of related posts on logging to a database, though I’m not quite sure if that makes much sense wrt. aggregation from multiple logging sources: Reddit discussion here and related Github repo here; a Gist of a db logging handler here; and Stackoverflow discussions python logging to database and Is this postgresql logging handler correct?. ↩