How to configure alerting for critical QuestDB errors based on logs?

Question

I'm configuring monitoring for my QuestDB instance and I'd like to have alerts when a critical error in the database. How do I do that?

Answer 1

QuestDB has two types of error log levels: normal (recoverable) errors and critical errors. Normal errors have " E " text in the message, e.g.:

2023-11-30T18:06:11.394326Z E i.q.c.l.t.LineTcpConnectionContext [22] could not parse measurement, NONE at 28, line (may be mangled due to partial parsing): 'up ..=4.0 631170000000000000'

These errors are usually due to invalid user input, e.g. incorrect SQL.

Critical level messages have " C " test in the message. Example:

2023-11-30T18:08:30.824047Z C i.q.c.w.s.TableSequencerImpl could not apply structure change to WAL table sequencer [table=testDodgyAddColumDoesNotChangeMetadata~1, error=[0] invalid alter table command [code=0]]

These errors are non-recoverable, so that you should monitor for critical-level messages in your alerting system.