Sentry shows a large amount of ResizeObserver errors from the same source

Description

A very large percentage of errors reported in the frontend belong to code_review_workflow, and many of those are related to ResizeObserver loop limit exceeded, I think it’s worth having a deep look as they may stem from a single source:

Error: https://new-sentry.gitlab.net/organizations/gitlab/issues/?project=4&query=is%3Aunresolved+feature_category%3Acode_review_workflow&statsPeriod=24h

Proposal

Investigate where if these errors come from a single source and patch the issue, so our error rate is improved.

Related data

Our sentry data shows this is a very common error, with over 100k events per day.

https://new-sentry.gitlab.net/organizations/gitlab/dashboard/3/

https://new-sentry.gitlab.net/organizations/gitlab/issues/?project=4&query=is%3Aunresolved+feature_category%3Acode_review_workflow&statsPeriod=14d

added Sentry WorkingGroupFrontendObservability bugtransient groupcode review labels

added typebug label

added devopscreate sectiondev labels

@slashmanov perhaps this is something you are having a look at? I'll assign to you.

cc @andr3

assigned to @slashmanov

changed title from Sentry shows a large amount of ResizeObserver errors from the same sources to Sentry shows a large amount of ResizeObserver errors from the same source

mentioned in issue #394700

added frontend label

changed milestone to %15.11

set weight to 3

mentioned in issue #386703 (closed)

mentioned in issue #395795 (closed)

set weight to 1

added Deliverable label

Setting health status to on track as the milestone has just begun.

Issue participants are welcome to override this by setting the health status to another value.

changed health status to on track

This issue is scheduled for completion in this milestone but is not yet in development. Changing health status to 'needs attention'.

Issue participants are welcome to override this by setting the health status to another value.

changed health status to needs attention

This issue is scheduled for completion in this milestone but is in an early development stage. Changing health status to 'at risk'.

Issue participants are welcome to override this by setting the health status to another value.

changed health status to at risk

added workflowin dev label

changed milestone to %16.0

added missed-deliverable missed:15.11 labels

removed workflowin dev label

We're going to remove ResizeObserver from the FileTree. While I was working on it I've noticed that these errors are still there even with the changes applied. That means that the bug is caused primarily by a 3rd party library (vue-virtual-scroller).

I am able to reproduce the error locally by changing zoom very quickly. The error is most definitely caused by DynamicScroller.vue (part of our "fork" of a 3rd party). Unfortunately it's very hard to understand what's going on because there's no error stack:

Previously we thought this was caused by the FileTree virtual scrolling so that should be out of the question now.

I am going to unassign myself for now because I don't have the capacity to dig even deeper on this. Anyone's free to pick it up, especially if you have experience with virtual scrolling.

/cc @andr3

@slashmanov thanks for the insight.

@iamphill would you say we're close to finding a solution for this? Should we schedule a spike for 16.1 at least? WDYT? Anything else we need to do before?

There is a fix inside of !119753 (merged) that fixes the 3rd party library that we either need to look into why it is happening or push the changes upstream. Though at least this change should fix the errors for now

@iamphill great, thanks for the update. Let's at least pursue mitigating the flooding of sentry with this into 16.1

@iamphill can you please help us figure out if this is still a concern? In sentry it still shows quite a good deal of events, still.

/cc @thomasrandolph

With !119753 (merged) being merged and deployed it looks like the number of events has dropped. Not sure what else can be done.

I think this issue is solved! We have sharp decrease of these events in the last 24 hours:

Search: https://new-sentry.gitlab.net/organizations/gitlab/issues/?environment=gprd&project=4&query=is%3Aunresolved+feature_category%3Acode_review_workflow+ResizeObserver&referrer=issue-list&statsPeriod=24h

I think we could move this to workflowverification and keep it under observation for a day or more and close this if the trend stays the same

@ayeung If you happen to be monitoring new-sentry's resource consumption you may also see decrease now that we are processing fewer events.

Aha, we just checked too soon! Glad to hear it's tapering off.

@andr3 I think I should unassign myself since I had nothing to with resolving this other than watching the sentry sparkcharts go to zero like

Thanks, @iamphill!

I'll close this issue as done now that errors are not reported anymore. Thanks all for looking into this!

unassigned @slashmanov

Setting health status to on track as the milestone has just begun.

Issue participants are welcome to override this by setting the health status to another value.

changed health status to on track

mentioned in merge request !115867 (merged)

This issue is scheduled for completion in this milestone but doesn't have an assignee. Changing health status to 'needs attention'.

Issue participants are welcome to override this by setting the health status to another value.

changed health status to needs attention

mentioned in merge request !115173 (merged)

This issue is scheduled for completion in this milestone but doesn't have an assignee. Changing health status to 'at risk'.

Issue participants are welcome to override this by setting the health status to another value.

changed health status to at risk

changed milestone to %16.1

Setting health status to on track as the milestone has just begun.

Issue participants are welcome to override this by setting the health status to another value.

changed health status to on track

assigned to @thomasrandolph

added workflowverification label

marked this issue as related to #412874 (closed)

unassigned @thomasrandolph

changed the description

closed

@mrincon, the workflow label was automatically updated to workflowcomplete because you closed the issue while in workflowverification.

If this is not the correct label, please update.

To avoid this message, update the workflow label as you close the issue.

added workflowcomplete label and removed workflowverification label

mentioned in issue gitlab-org/quality/triage-reports#12804 (closed)