Hello all,
We are looking to move our company's looker user to a Redshift Serverless database instead of the current provisioned cluster database that Redshift natively uses.
We've prepared all of the necessary items within this serverless cluster (sharing data from our main cluster to serverless, created looker users, verified data exists).
The only issue we've run into so far has been in looker's attempts to kill queries. This is an imperative operation that we need available for our looker cluster as Redshift Serverless only charges our account by the time a query is ran. If we are unable to kill a query that is not necessary, we will be getting charged for the duration of the query.
After doing some digging on the queries ran when testing a new looker database connection, we can see the error lies within this query:
```
SELECT pid as id, duration / 1000000 as elapsed_time, query as stmt FROM stv_recents WHERE status='Running' AND user_name = '#{username}'
```
According to redshift docs (https://docs.aws.amazon.com/redshift/latest/mgmt/serverless-monitoring.html), the STV_RECENTS table is nonexistent within Redshift serverless. Nevertheless, there are certainly other ways for looker to complete cancellation requests (like, 'select pg_terminate_backend(<pid>);')
My request is for a new feature that will allow for a connection to Amazon Redshift Serverless with a slight few tweaks in how the connection handles killing queries.
Thanks!
Reid Williams
(PII Removed by Staff)
LinkedIn
Twitter
