Looker Studio Connection to Cloud SQL (PostgreSQL) Fails Persistently (Error b33ffb08)

Google Cloud Database
1

Hello everyone,

I’m facing a persistent authentication error when trying to connect Looker Studio to my Cloud SQL for PostgreSQL instance. I have already exhausted all standard and advanced troubleshooting steps, and I’m beginning to suspect a platform-level issue.

I would be very grateful for any insight or help.

Environment Details:

  • Project ID: (PII Removed by Staff)

  • Cloud SQL Instance Connection Name: (PII Removed by Staff)

  • Database Engine: PostgreSQL 17

  • Database Name: estoque

  • Looker Connector: Using the native “PostgreSQL” connector.

Error Received: A generic error message: “An error occurred, and it was not possible to complete your request.” with various error codes, including b33ffb08, db26bd7a, and 82ba48d2.

Troubleshooting Steps Already Performed:

We have systematically checked and eliminated all common causes of failure:

  1. Correct Connector: We have confirmed we are using the PostgreSQL connector, not MySQL.

  2. Correct Connection Method: We are using the URL JDBC tab, which is the correct method for connecting to Cloud SQL’s internal network via the Auth Proxy. The URL is formatted correctly: jdbc:postgresql://google/estoque?socketFactory=om.google.cloud.sqcl.postgres.SocketFactory&socketFactoryArg=(PII Removed by Staff)east4:instancia-financeiro-prod

  3. Cloud SQL Admin API: We have verified that the Cloud SQL Admin API (sqladmin.googleapis.com) is ENABLED for this project.

  4. Firewall / Authorized Networks: To completely eliminate the firewall as the cause, we added an authorized network rule for 0.0.0.0/0. The connection error persisted even with the firewall completely open for testing.

  5. Credentials (Exhaustively Verified): This has been our main focus.

    • We initially corrected a typo in the username from postgre to the correct default superuser, postgres.

    • The password for the postgres user was reset multiple times to ensure it was correct.

    • Crucially, we created a brand new, clean user (looker_user) with a new, strong password. The connection attempt failed with this new user as well.

Conclusion: Since the connection fails even with a brand new user, with a completely open firewall, and with the correct APIs and connection strings, this does not appear to be a user configuration error.

Could there be an internal issue with this specific Cloud SQL instance or a project-level setting that is preventing authentication from Looker Studio?

Thank you for any help you can provide.

2

Hi Rehael_Henriques,

Looker Studio currently supports PostgreSQL versions up to 14, and you’re using PostgreSQL 17, which is not officially supported. This mismatch can lead to unpredictable behavior, including generic authentication failures like the ones you’ve encountered. You might also consider testing the connection via a public IP with SSL enabled, adding Looker Studio’s IP ranges (such as 142.251.74.0/23) to your authorized networks. Ensure that your service account has the Cloud SQL Client role assigned. If it’s not assigned, kindly add it.

Additionally, you can check this case in Stack Overflow, as it might be related to your issue and could provide helpful insights.

If you try this workaround and the issue still persists, you may consider reaching out to Google Cloud Support for further assistance.

1 Like