none
Intermitent slow queries RRS feed

  • Question

  • We've just ported a locally stored postgres database from on-prem use to Azure an managed postgresql instance.

    Every once in a while, (1-2 days frequency) for a few minutes something happends to the database that even a simple query takes well over 30 seconds, or just times out and I don't know what to do unless do some refreshes... Usually it just resumes working, without intervention and I have no Idea what causes it... or what to look for in the logs.

    The CPU is under 20% utilization, and no spikes in any of the activity monitored around the time of the incidents.

    The memory seems to stay quite high, at 60% but there are no indications to be worse at the time of the incidents, and it the rest of the time, it just works...

    The logs are not that explicit either

    All I get is: "could not send data to client: An existing connection was forcibly closed by the remote host."

    Not sure what I should look for. Probably some simple setting I forgot to check.

    Thanks,

    Cris

    Thursday, September 20, 2018 6:19 PM

All replies

  • Hi Finebytes,

    Do you have timestamps for when these specific occurrences happen?

    Do you monitor any aspects of your PostgreSQL instance?

    Let's investigate this. Regards, ~Mike

    Friday, September 21, 2018 7:00 PM
    Moderator
  • Can you please provide your instance name, as I can have your specific deployment investigated for any operational issues. Another deployment was restarted and this resolved a similar issue another customer was experiencing.

    In fact, you may not be experiencing issues going forward. Please confirm. Thank you!

    Friday, September 28, 2018 3:51 PM
    Moderator