1788228 - sentry events include wrong information

Status: RESOLVED FIXED

(Socorro :: General, defect, P1)

Description

[Will Kahn-Greene [:willkg] ET needinfo? me]

This Sentry event is for "timeout" in pymemcache presumably on /__lbheartbeat__:

https://sentry.io/organizations/mozilla/issues/3546766253/

However, there are several instances in that issue where the transaction information (url, http headers, query string, transaction, etc) for the issue is for a SuperSearch request which has a different stack.

Seems like when the timeout happens, Sentry client is picking up information from a different thread or something like that.

Is the Sentry client set up correctly? Are we hitting a Sentry client bug?

Comment 1

[Will Kahn-Greene [:willkg] ET needinfo? me]

We're using sentry-sdk 1.9.0 and they're up to 1.9.5 now. It's worth updating as a first pass.

Comment 2

[Will Kahn-Greene [:willkg] ET needinfo? me]

Making this a P1 to fix soon because it's really unhelpful to have junk data in Sentry.

Comment 3

[Will Kahn-Greene [:willkg] ET needinfo? me]

Attachment: pr 6160: bug 1788228: update sentry to 1.9.5

Comment 4

[Will Kahn-Greene [:willkg] ET needinfo? me]

willkg merged PR #6160: "bug 1788228: update sentry to 1.9.5" in 049a8be.

Comment 5

[Will Kahn-Greene [:willkg] ET needinfo? me]

We'll have to see what that does in production because we're not getting the error in stage.

Comment 6

[Will Kahn-Greene [:willkg] ET needinfo? me]

I pushed this to prod just now in bug #1788726. I'll keep this open for a week or so to see what happens.

Comment 7

[Will Kahn-Greene [:willkg] ET needinfo? me]

I don't think I've seen more evidence of this. I'm not sure why it was happening and I'm not sure whether it's really fixed, but I'm going to mark it FIXED now anyhow.