Hello Georges-André Janett,
Thank you for reporting this.
The issue you were experiencing where the Azure Container Registry (ACR) agent pool gets stuck in the "Updating" state typically occurs when an update operation (such as resizing or scaling the pool) fails to complete and the resource becomes locked in a transitional state. This prevents further operations like updating, deleting, or scaling the agent pool, and results in errors like:
(RegistryStatusConflict) Cannot perform the operation as the resource 'devops' is in 'Updating' state.
What was done to fix this?
In your case, we resolved the issue through a backend remediation. I've manually reset the provisioning state of the agent pool using internal tools. Once that was done, the az acr agentpool update
command succeeded.
The command that was used to scale the pool to zero was-
az acr agentpool update -n devops -r crimdpreprod -g rg-devops-preprod --count 0
After the backend state was corrected, the command executed successfully, confirming that the agent pool had exited the "Updating" state and was back in a healthy "Succeeded" state.
The stuck state was due to an incomplete or failed internal update that did not finish properly. This left the agent pool in a locked provisioning state. These situations require manual intervention as Azure’s automatic remediation may not always recover them promptly.
It would be great if you could kindly mark the answer as accepted so that anyone having similar query on MS QnA forum can refer to this and be benefited from it. Thanks