RE: Failed transaction statistics to measure the logical replication progress

Поиск
Список
Период
Сортировка
От osumi.takamichi@fujitsu.com
Тема RE: Failed transaction statistics to measure the logical replication progress
Дата
Msg-id TYCPR01MB83736371B66F90C71365ED11ED7B9@TYCPR01MB8373.jpnprd01.prod.outlook.com
обсуждение исходный текст
Ответ на Re: Failed transaction statistics to measure the logical replication progress  (Kyotaro Horiguchi <horikyota.ntt@gmail.com>)
Список pgsql-hackers
Friday, December 17, 2021 2:03 PM Kyotaro Horiguchi <horikyota.ntt@gmail.com> wrote:
> It sends stats packets at every commit-like operation on apply workers.  The
> current pgstat is so smart that it refrain from sending stats packets at too high
> frequency.  We already suffer frequent stats packets so apply workers need to
> bahave the same way.
>
> That is, the new stats numbers are once accumulated locally then the
> accumulated numbers are sent to stats collector by pgstat_report_stat.
Hi, Horiguchi-san.

I felt your point is absolutely right !
Updated the patch to address your concern.

Best Regards,
    Takamichi Osumi


Вложения

В списке pgsql-hackers по дате отправления: