We had previously noticed behavior where Sentry would flag commits for issues that couldn't possibly be related, for example due to being in a release that was shipped before the commit was landed. Since then, we've improved our integration to leverage semver versions, and this behavior hasn't improved.
Expected Behavior
New issue is detected in a release much older than the current latest release in the environment (eg, a four week old release to stable)
Suspect commits should find the latest commit in the stack trace based on that version
barring that (missing commit data), it should at least only search for commits merged before the release time of the release that the issue was first seen in
Actual Behavior
New issue is detected in a release much older than the current latest release in the environment (eg, a four week old release to stable)
Suspect commits flags a commit landed yesterday, which cannot possibly have caused the issue
both from a release train perspective, but also a time-based perspective
Side Question
Would setting up commit association improve this behavior, or is that not functionality that suspect commits hooks up to yet? Our instance does have Github integration setup.
We had previously noticed behavior where Sentry would flag commits for issues that couldn't possibly be related, for example due to being in a release that was shipped before the commit was landed. Since then, we've improved our integration to leverage semver versions, and this behavior hasn't improved.
Expected Behavior
stable
)Actual Behavior
stable
)Side Question
Would setting up commit association improve this behavior, or is that not functionality that suspect commits hooks up to yet? Our instance does have Github integration setup.
PS: Following up internally with examples