Hello friends, we are stating to see a lot of

sql: transaction has already been committed or rolled back

errors in our logs which we suspect is causing the hydra

/health/ready

endpoint to fail. Have any of you people seen this before? we are using Hydra v. 1.10.7 with a postgresql database.

can you share your code example

Well, it is Hydra that is logging the error, so I do not have any code to show. This is in the logs from running the hydra executable.

Check in the DB for connection error and try to restart hydra and7or the db

Restarting fixes the issue temporarily but after some time it returns. The database errors all seem to happen when calling the

/oauth2/token

endpoint.

In that case you might need to fix up your DSN: https://www.ory.sh/docs/ecosystem/deployment/

I think we figured out the issue. Circling back to our DSN as you suggested we took a look at the database and saw that it was depleted of resources. Allocating more resources to the database seems to have fixed the issue. Thank you all for the helpful suggestions 🙏