Initial post: 9 May 2021, 20:11 BST
Last update: 1 Jun 2021, 10:46 BST

On the 8th of May many users started reporting issues with communicating against Sandbox orgs hosted on CSXXX Salesforce instances.

We have contacted our partners on the 10th of May about this issue.

On 21st of May, our partners published this RCA on the issue. Gearset has informed Salesforce support to clarify that the RCA is not the issue we reported. The timings and affected resources do not match our observations.

Since 03:41 AM of 14th of May, our monitoring of the issue demonstrates that the issue is no longer happening and would be interested to hear if users are still seeing this. The API communication started working again, consistently, without changes to the app made by Gearset.

What's the issue?

The error may be presented in Gearset as:

  • An unknown error occurred.

  • Determining API versions - An unknown error occurred.

  • Gearset could not contact Salesforce with this login.

  • An internal server error has occurred for unit tests.

Gearset was unable to successfully communicate with the API all the time. We understand the issue is intermittent and the API communication does not always fail. The same exact same query or request to our partner succeeds in some cases but may fail on a subsequent identical request.

When it did fail, the error An internal server error has occurred is returned, and that is sometimes exposed in the Gearset app UI.

What orgs does it affect?

All failure are reported to be sandbox orgs on the CSXXX instances.

We do not think this issue is related to the upgrade to Summer '21.

Is there a workaround with Gearset?

As the issue is reported as intermittent, other users have been able to succeed trying a bit later.

If you are running a comparison and it's not an automated job - you can switch to using username/password option to connect the org.

What is Gearset doing?

Based on our testing, we are somewhat confident this is not an issue caused by Gearset's configuration.

We have added some mitigation to the retry logic that would reduce impact of problem - that was shipped on the 11th of 12th of May.

We were also monitoring for any updates.

Did this answer your question?