Horton

Re: Add jsonlog log_destination for JSON server logs

Michael Paquier <michael@paquier.xyz>

View thread
From: Michael Paquier <michael@paquier.xyz>
To: Sehrope Sarkuni <sehrope@jackdb.com>
Cc: Pg Hackers <pgsql-hackers@postgresql.org>, david@fetter.org
Date: 2021-09-10T06:56:18Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Introduce log_destination=jsonlog

  2. Refactor set of routines specific to elog.c

  3. Fix compilation warning in syslogger.c

  4. Refactor per-destination file rotation in logging collector

  5. Refactor output file handling when forking syslogger under EXEC_BACKEND

  6. Add regression tests for csvlog with the logging collector

  7. Refactor the syslogger pipe protocol to use a bitmask for its options

Attachments

On Fri, Sep 10, 2021 at 01:07:00PM +0900, Michael Paquier wrote:
> Forking a bit this thread while looking at 0002 that adds new tests
> for csvlog.  While I agree that it would be useful to have more
> coverage with the syslogger message chunk protocol in this area, I
> think that having a separate test is a waste of resources.  Creating a
> new node is not cheap either, and this adds more wait phases, making
> the tests take longer.  It would be much better to extend
> 004_logrotate.pl and update it to use log_destination = 'stderr,
> csvlog', to minimize the number of nodes we create as well as the
> additional amount of time we'd spend for those tests.  Plugging in
> JSON into that would not be complicated either once we have in place a
> set of small routines that limit the code duplication between the
> checks for each log destination type.

And this part leads me to the attached, where the addition of the JSON
format would result in the addition of a couple of lines.
--
Michael