Hello everyone, we're collecting snapshots of the statistics (pg_stat, pg_statio etc.) regularly so that we can analyze the performance, detect trends etc. We need to identify whether the stats were reset between the snapshots, because then the collected data are useless (and the scripts that do the analysis are quite confused). Thanks to pg_postmaster_start_time it's quite simple to find out whether the postmaster was restarted (before 8.3 that may reset the stats if stats_reset_on_server_start=true). But is there a way to find out whether the statistics were reset using pg_stat_reset_* functions? E.g. 'timestamp of the last reset' or something like that? Currently I'm comparing some of the cumulative values (e.g. n_commits) and if the newer version is lower than the old one, it's considered as a proof that the stats were reset. But I don't like this solution and it does not detect resets of statistics for a single table, function or a bgwriter :-( But is there a better way? regards Tomas -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
