Re: [PROPOSAL] timestamp informations to pg_stat_statements

2018-04-04 Thread legrand legrand
> No, the entries are not removed randomly. We track "usage" for each > entry (essentially +1 for each time the query got executed, with a decay > factor applied on each eviction (and we evict 5% at a time). OK I apologize, I hadn't looked in source code in details, and effectively the "Usage" alg

Re: [PROPOSAL] timestamp informations to pg_stat_statements

2018-03-10 Thread Tomas Vondra
On 03/10/2018 04:43 PM, legrand legrand wrote: > +1 > Having the time of first occurence of a statement is very usefull > for trouble shouting, it permits for exemple to retrieve the order of > operations in some complex cases (and thoses informations aren't > taken by any third party collecting

Re: [PROPOSAL] timestamp informations to pg_stat_statements

2018-03-10 Thread legrand legrand
+1 Having the time of first occurence of a statement is very usefull for trouble shouting, it permits for exemple to retrieve the order of operations in some complex cases (and thoses informations aren't taken by any third party collecting tool, that will only be able to provide a time range of occ