How can I connect a Cloud SQL instance to my Cloud Run V2 push-Work Pool?

[AmagicalFishy]

This would be the equivalent of deploying the job using w/ gcloud run deploy ... --add-cloudsql-instances ${INSTANCE}. There's a space to enter the VPC connector (equivalent to the --vpc-connector=${CONNECTOR} flag, so I assume adding a Cloud SQL instance isn't far off.

Now, I might be crazy, but the option popped up for one of my newly-created workpools. Now I can't get it to come up for the life of me (I'm creating work pools via the UI). I even made deployments to that pool.

Note: This is a push work pool, so I do not manually create workers; they're created automatically.

[parkedwards]

@AmagicalFishy doesn't look like we expose this as a UI form field currently, but you should be able to modify the push pool's job configuration in the Advanced tab to add it in. this spec is the source of truth for the payload that gets sent to Cloud Run, which defines the CR job

I believe the --add-cloudsql-instances corresponds to this argument on CR V2's API
https://cloud.google.com/run/docs/reference/rpc/google.cloud.run.v2#google.cloud.run.v2.CloudSqlInstance

if I'm translating the docs correctly, I think this might be the right spot to drop in your arg

[AmagicalFishy]

Ah! Thanks, you're a lifesaver! I was looking through this and thus structuring my JSON pretty incorrectly.

Man, I absolutely swear that one of my pools had this option and I just clicked and selected my SQL instance. 🤔 Anyway that worked, thanks again

Edit:
That actually didn't work! What did work was adding:

    "cloudsql_instances": [
      "bdv-tech:us-east1:bdv"
    ]

... as an entry to the job_configuration key 😉

[Nonso-M]

I tried adding that as a key but still have the connectivity issue. Did you leave the volumes mentioned above?