Retool starting our Databricks serverless SQL without any query being processed

Hey there!

We are using the Databricks connector to connect our Retool app with our datalake through a Serverless SQL datawarehouse (which spins up among any query and has a 5min sleep time) and we are running into a situation where Retool is spinning it up every 30min without executing any query which is increasing our Databricks bill considerably.

We have a 1 day cache policy (maximum allowed) for all the queries that are executed to Databricks but still we can't understand why the cluster is being spinned up among non activity in Retool.

Below you can see a screenshot for a time frame showing that no queries were processed but the cluster is being spinned up with a repetitive behavior:

Any help regarding how can we mitigate this behavior would be greatly appreciated.

Unfortunately we don't have logs visibility in terms of what API calls Retool is executing against Databricks cluster.

Thanks a lot!

Patricio

Hello @pbenvenuto!

Sorry for this issue, very odd behavior.

From the Retool side of things, you should be able to see a run log of when queries are running. So you can view if/when those queries are being called.

It sounds like you are not running any queries manually, so I am not sure what would be triggering a request to your databricks server other than a query :thinking: I can check with our core data team and see if this is a known issue.

My guess is that they will want to get the logs from the database server side of things to see what is coming in and where it is coming in from :sweat_smile: