SQL time out on panels

I’m using the SQL connector to a Microsoft SQL server. The graphs are working fine and connections are good. I have noticed when the server is bogged down or a query is running slow, it appears that their is a default 30 second time out somewhere. It says sql_timeout and I get the red icon in the top left of the panel with the error.

If I decrease the timeframe or wait until server is running better, it has daily jobs that slow it down, and refresh the page, it works again fine.

I checked with DataGrip, the time out is not from the database, queries can run there for as long as needed.

Is this a parameter somewhere for the SQL driver? Is the error correct or misleading, is it a time out on the frontend not waiting long enough or really with the SQL connection?

I want a timeout, but closer to 2 minutes, or 120 seconds.

I tried the query string parameter thing I read in these forums and it didn’t help.

Any suggestions?

I found this, maybe just not supported.

I know I should push this data to it’s own reporting server in maybe Influx, but just trying to get some initial dashboards working quickly with existing database I have and the tool is great for that, just need more time!

The error actually says…

read tcp ipaddress i/o timeout

Configurable timeouts are still not implemented and are not high up our backlog. The core team would accept a PR but will not be working on this in the near future.

What about for short term if the timeout for SQL and Postgres was changed to default to 60 instead of 30, since they aren’t time series and can be slower? That should be a quicker than configruable.

If that’s still low priority, would you think core team would take that PR? That would be a super fast one.

good question - not sure. I suppose so.

Here is an example of adding the config option for Prometheus: