I updated a dashboard. Then I published dashboard (with update-dashboard-published-version). I check for error if CREATION_IN_PROGRESS then will retry to publish again later. But if it’s any other status/error will not attempt to retry. I see the that a new version of dashboard was created with CREATION_FAILED status and the error is “The processing of the dashboard failed because of concurrent updates. Please poll for the status of the dashboard before making other updates.”
What are the concurrent updates it is referring to? How can I recover from this failed state? I need the new version published.
Hello @alltej , is there a chance you attempted to update a dashboard that someone else had updated and you somehow didn’t have the most recent version? I think this is pretty well handled by Quicksight but I suppose it could be a possibility. You could attempt to create a fresh analysis off of the published dashboard, make those changes again, and attempt to publish it again.
If you continue having issues, I saw this other question in the community that was similar and Peter suggested logging a ticket with the technical team to debug the issue. That could be a useful path if you continue to experience this error.
There is no chance that someone else could be updating because there is only one process that is doing deploy and there is really no other manual or user updates. If I rerun it again, it will update and publish successfully. But I think this is a bug during the update because it should retry again internally and not a user initiated retry. The new version should stay in CREATION_IN_PROGRESS if there are concurrent updates because there is no way to publish it if it is already in CREATION_FAILED. This type of error is INTERNAL_ERROR so it is not something a user code/config error. During publish, I try to retry later if status is in CREATION_IN_PROGRESS. But if it is already failed, I have to start over again in the update and publish process.
@alltej there is a possibility you are running into a bug on this issue. What are you using to update the Dashboard, CLI or SDK? That might help direct me a bit in finding a solution for this issue.
Hi @alltej , if this is all working for you now after rerunning the command I will move this question forward. I would recommend if this occurs for you again you could file a claim with the technical support team at AWS to sort out a solution for the bug. I apologize for not having a proposed solution, but this seems to be the proper course of action.