Slack Integration: error event links to latest error instead of actual timestamped error instance

[tlowande]

Environment

SaaS (https://sentry.io/)

Version

No response

Steps to Reproduce

"@sentry/node": "5.30.0"

1. have a sentry error grouped by message. These errors are built with Sentry.captureEvent from @sentry/node. They have the same message but we add different values under extras: {splat}
2. have these errors reported in a slack channel through slack integration
3. have the error triggering more than once (triggering for different reasons - meaning the values under splat are different, but the error message is the same)

Expected Result

if the error posted in slack is part of the same group (grouped by message) the link of each instance of that error in the slack message is unique and leads to that instance and timestamp of that error instead of the most recent event of that group

Actual Result

if the error posted in slack is part of the same group (grouped by message) the link of both instances of that error in the slack message is the same and leads to the most recent timestamp event of that group. So two different events of that error that happened in 2 different timestamps have the same link which is connected to the latest instance of that error.

┆Issue is synchronized with this Jira Improvement by Unito

[getsentry-release]

Routing to @getsentry/ecosystem for triage. ⏲️

[leeandher]

Hi @tlowande, thanks for bringing this issue up. The notifications work and link as intended, though as you said, the events that happened to trigger the alerts are not directly linked to in the notification, we just link the generic issue as a whole.

I was able to reproduce it on my end and I agree that it seems like a nice feature we should definitely look at implementing, so I'll be adding it as a ticket to our team's backlog to prioritize. Thanks again!

[dnvarun]

Stumbled upon this issue while googling regarding this behavior. Is there a way to link to actual timestamped error from sentry dashboard? I looked at the alert settings but was not able to find anything related.

[rdsedmundo]

There's not a single dev that enters the team which doesn't misinterpret this, and even myself that use Sentry for almost a decade have to constantly remind myself of this behavior as it's totally misleading. I'm surprised not more people are complaining about this / or that it has not been resolved yet, it should be an option in the integration at the least.

You click on something that says error with User X (request A), and then you enter the page and it's actually referencing User Y (request B).

Tagging rachrwang as I saw you work in Product fro Sentry and answered a similar question a couple months ago.

[getsantry]

Routing to @getsentry/product-owners-settings-integrations for triage ⏲️

[Dhrumil-Sentry]

@rdsedmundo thanks for this feedback

it should be an option in the integration at the least.

This is probably the right path here. We're adding this to our backlog and will keep the issue updated as we have more details

[mjhoy]

Curious if there are any updates on this. This causes our team quite a bit of confusion, and can lead to missing important errors.

[sidraval]

I would love to see this implemented – we use Sentry as part of critical monitoring, and it's wildly confusing that clicking a particular error in slack may not take you to that particular instance in Sentry.

[leedongwei]

This issue is still in the backlog. Sorry!