Horton

RE: Re: Log a sample of transactions

Kuroda, Hayato <kuroda.hayato@jp.fujitsu.com>

View thread
From: "Kuroda, Hayato" <kuroda.hayato@jp.fujitsu.com>
To: 'David Steele' <david@pgmasters.net>, Adrien NAYRAT <adrien.nayrat@anayrat.info>
Cc: Andres Freund <andres@anarazel.de>, "'pgsql-hackers@lists.postgresql.org'" <pgsql-hackers@lists.postgresql.org>
Date: 2019-03-26T00:54:48Z
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. Tweak docs for log_statement_sample_rate

Dear David,

I have a will and already read the patch, but I thought it's not my turn.
Sorry.


Adrien,

>  I did not find any test for log_min_duration that could help me. LCOV indicate
> there is no tests on this part (look check_log_duration()):
> https://coverage.postgresql.org/src/backend/tcop/postgres.c.gcov.html

I understand the unnecessarily of some test case. It's OK.

Finally, how do you think about the deviation of randomness?
If this parameter is set very low, nobody may be output because of the deviation.
we can avoid this phenomenon by counting up internal parameter for each transactions and output to log file if the parameter becomes  more than 1.

After consideration for this case and rebasing, I think this patch is enough.
Do I have to measure the change of throughput?


Best Regards,
Hayato Kuroda
Fujitsu LIMITED