Is there a way to automatically remove issues that are x days old?
Have a job like this…
Kind of hard to have if you are running cloud hosting
@Mackelito Is your motivation for this to delete sensitive information or to clean up the issues list visually? There’s an auto-resolve setting for the latter. And we generally don’t store events older than 90 days if you’re concerned about privacy.
Clean up due to GDPR and our own customer agreements.
You can delete individual issues if you need to, but you cannot configure the global retention. Events are automatically deleted after 90 days. I would recommend sending anonymized PII if this is a concern (e.g. send the user’s ID instead of their email address).
I see… is that something that you perhaps would be willing to add as a feature?
It would greatly simplify things for us if there where a global way to resolve this…
Perhaps some configuration allowing for custom fields to be cleared after x time?
Of course we can just do a search for the email adress and remove issues… but it’s a manual work not something I trust to be handled at all times
It might be something we offer in the future, but we don’t offer it today as the complexity is extremely high and the use cases for it generally can be worked around.
So just a recap…
- Events are purged after 90 days
- Issues stick around
Issues stick around generally speaking until no events remain related to the issue.
We link to Sentry issues in bugs that we file in Jira.
This has been awesome for our team finding and fixing bugs when reported, and also has been super helpful when bugs get reopened due to regressions.
We are running into a problem however which is our older Jira tickets have links to Sentry issues that no longer exist (through no action of ours).
Is there a way to “protect” a Sentry issue from being deleted? Like a “share” or “bookmark”, etc.
Unfortunately not. It’s something that technically wouldn’t be too hard to keep around (at least in some form), and we’ve certainly talked about it before. The challenge right now is events operate on a TTL schedule to optimize deletions. That means that keeping any event past e.g. 90 days is actually pretty tricky. We’d have to write them into a different storage to achieve that, which would mean they’re only available in some odd capacity. Keeping the issue around tho is fairly trivial, its just not super useful without a sample event to go along with it.
Thank you for the speedy response and the detailed explanation.
Our workaround at present is to continue to include the Sentry link in the Jira ticket and to also copy/paste the exception from the Sentry issue into the Jira ticket.
Let me know if you have other ideas or have heard of other approaches customers are taking to accomplish the same goal.