Open coury-clark opened 2 years ago
There are some fields in the insight series resolver that map to counts in a postgres database. This table can be quite large as it represents queued (or completed / errored) Code Insight queries. Postgres queries that are just asking for a count can be quite slow, and it seems to now be a dominating factor when I poke around insights and inspect their traces.
Heads up @joelkw @felixfbecker @vovakulikov @unclejustin - the "team/code-insights" label was applied to this issue.
There are some fields in the insight series resolver that map to counts in a postgres database. This table can be quite large as it represents queued (or completed / errored) Code Insight queries. Postgres queries that are just asking for a count can be quite slow, and it seems to now be a dominating factor when I poke around insights and inspect their traces.