401 errors due to JWT rejections

Updates timeline

1. Identified Aug 14, 02:23 AM UTC

   We have identified the root cause of newly refreshed JWTs being rejected by the API, resulting in HTTP 401 errors for affected sessions. We are working on a fix and will provide updates as things progress.

2. Identified Aug 14, 07:53 AM UTC

   Our teams are continuing to work on the fix and rollout is underway for selected customers. We will provide another update as we are applying the fix to all impacted users.

3. Identified Aug 18, 06:38 PM UTC

   Fixes for this issue are being sequentially rolled out. The impact of the issue is limited to a subset of new projects that can experience it upon some JWT renewals. We will update the status when all the fix rollouts have completed.

4. Identified Aug 20, 04:37 PM UTC

   As part of the incident remediation flow, we have separately identified "/lib/aarch64-linux-gnu/libc.so.6: version `GLIBC_2.34' not found" error between 11:15 - 16:24 UTC today affecting PostgREST. This has been fixed and we can see the error rates have subsided.

5. Identified Aug 21, 05:21 PM UTC

   We are continuing to test and rollout the fixes for intermittent HTTP 401 errors. In most cases, waiting and refreshing is successful. We will monitor eu-central-2 over the weekend and are preparing for the fix to be pushed to all regions starting Monday. Thank you for your patience while we've worked on multiple fixes for this issue.

6. Identified Aug 25, 06:41 PM UTC

   We are continuing to rollout the fixes for intermittent HTTP 401 errors. In most cases, waiting and refreshing is successful. This fix has been applied to the following regions: ap-east-1 ap-northeast-1 ap-northeast-2 ap-south-1 ap-southeast-1 ap-southeast-2 ca-central-1 eu-central-1 eu-central-2 eu-north-1 eu-west-2 eu-west-3 We will continue rolling out this fix to the remaining regions throughout this week and provide updates as more regions are successful.

7. Identified Aug 27, 05:17 PM UTC

   PostgREST 14.17 has now been rolled out to all regions. Some customers have reported that restarting their project after the rollout resolved this issue. To restart your project, go to the General settings page in the dashboard and select Restart project. Please contact our support team if the issue persists after a restart.

8. Identified Aug 28, 11:44 PM UTC

   We continue to monitor the fixes introduced by this latest rollout. If you are continuing to see these errors, some customers have reported that restarting their project after the rollout resolved this issue. To restart your project, go to the General settings page in the dashboard and select Restart project. Please contact our support team if the issue persists after a restart.

9. Identified Aug 31, 07:52 PM UTC

   We have seen reports that the JWT rejection issue persisted with 14.17, and due to <a href="https://status.supabase.com/incidents/bv4ntm4x0btf">some unintended performance side effects</a>, we have rolled back to Postgrest 14.5. The team continues to investigate the JWT issue and hopes to have a solution soon.

10. Identified Sep 2, 07:38 PM UTC

    A stale time cache has been identified as the cause of this issue. A targeted fix has been written and is undergoing internal testing. It will be rolled out when testing is complete. Thank you for your patience while we investigated this issue. We will update the status page when the fix is rolled out.

11. Identified Sep 7, 09:49 AM UTC

    Testing results have been positive. Following some additional checks today, we plan to expand the rollout. We’ll continue to monitor progress and provide further updates as they become available.

12. Identified Sep 11, 11:41 PM UTC

    We continue to monitor progress and are slowly updating each region. Thank you for your patience on this issue.

13. Identified Sep 17, 07:30 PM UTC

    To resolve this issue, our fleet requires a change to our deploy process. We have spent this week making these changes and will be implementing them throughout this weekend, starting on Friday, September 18. Once this change is implemented, we will release a new Supabase version that impacted users can upgrade to from their dashboard. Completing that upgrade will resolve this issue. We recognize that this is added work for our customers. This issue has been open for over a month, as we tried to make this change on each project's behalf. In an effort to resolve this sooner, we have opted to make this deployment process change and ask users to upgrade.

14. Identified Sep 23, 07:26 AM UTC

    We have rolled out the deployment process changes to eu-central-2 and continue to monitor. We will provide additional updates as we make this change to other regions. Once this change is implemented across all regions, we will release a new Supabase version that impacted users can upgrade to from their dashboard. Completing that upgrade will resolve this issue.

15. Identified Sep 25, 04:23 PM UTC

    The deployment process changes have been rolled out across all regions. We will monitor throughout the weekend. The new Supabase version will be released early next week. Once it is released, impacted users can upgrade to from their dashboard to resolve this issue. Thank you again for your patience while we worked through this. We will provide an update once the new version is available and users can upgrade.

16. Resolved Sep 29, 06:46 PM UTC

    The latest Supabase version has been released. All users impacted by this issue can now upgrade from the dashboard to resolve this issue. Since this new version is released, we will resolve this issue. If you have any questions about upgrading, please refer to our documentation for more information: https://supabase.com/docs/guides/platform/upgrading Thank you again for your patience while we worked on resolving this issue.