-
Notifications
You must be signed in to change notification settings - Fork 25
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Feature suggestion: ignore certain errors #117
Comments
I just spotted https://github.com/elixir-error-tracker/error-tracker/blob/main/guides/Getting%20Started.md#ignoring-errors, which is a nice alternative to UI ignores I think. Perhaps we don't need this issue anymore. I'll leave it open in case UI ignores are a thing you're interested in |
Hey @craigfurman, thanks for sharing your thoughts. I see how this could be a useful feature. I need to think about this. I got a few ideas but need to put them in order first. |
I would like to separate the concept of "ignoring" an error and "silencing" it.
As you mentioned, we already have the capability of ignoring errors. Silencing errors is not possible yet. But it is something we would like to do. The way I see it the silenced state is orthogonal to the resolved/unresolved states.
Silenced errors would still be recorded, but would not trigger telemetry events and thus won't trigger telemetry integrations either. Silencing and de-silencing an error would be done directly from the UI. |
Pull request #130 implements this functionality. I've chosen |
Looks great! Thanks for thinking about and implementing this |
👋 My team recently started using this project, and we're definitely enjoying it so far. One UX issue we do have though, is that our workflow can't quite be expressed using the current status values (resolved and unresolved). I think that another possible state, "ignored", with the ability to filter the errors page to omit these, would solve this for us. A bit more detail on our workflow below 👇
We integrate with the telemetry to notify ourselves when a new or regressed (resolved -> unresolved) error occurs. We don't notify on subsequent instances of known errors. Zooming out, we use error-tracker to quickly alert us to new kinds of error, for triage, in conjunction with metrics-based alerting on a moving-average error ratio (e.g. "alert when >1% of HTTP responses on an endpoint have status 5XX in the last 30 minutes"). We believe balances MTTD with a workable signal:noise ratio.
As such, we tend to leave error-tracker entries related to transient, retryable errors (e.g. network flakes) open. This isn't because we don't care about them as such, but rather that a very low ratio of such errors, below some low percentage threshold, doesn't meet the urgency criteria to alert. Leaving the error-tracker entry open (unresolved) results in the notification behaviour we want (i.e. not notifying us).
I'd be interested in marking such entries as "ignored" in order to de-clutter the front page of error-tracker, which would make triaging of new entries easier.
Is this something you'd be interested in as a maintainer? I'd be happy to send a PR if so.
Cheers!
Craig
The text was updated successfully, but these errors were encountered: