Sometimes when a problem happens, it causes errors to be logged from various different places in our code, possibly even from different projects (e.g., a network connectivity issue that affects multiple microservices). The errors are semantically distinct and each of them could also potentially occur in isolation, so merging the issues together isn’t appropriate.
But what would be great would be a way to group them together temporally. “Incident” seems like the right name for what I’m thinking about. Basically: “The events from time X to time Y for issues A, B, and C are all a single incident.”
An incident would have its own comment stream and perhaps would support integrations with task managers. The incident would be reachable from the issue’s details page.
Marking an incident resolved would resolve all the associated issues, though any errors that continued to occur after the end of the incident period would still remain unresolved (subject to any “ignore until it happens N times” rules).
Being able to reason about and track issues with the same underlying cause would be super helpful when it comes time to write postmortem reports, etc.
The temporal aspect of this is key: a single issue could be a member of multiple incidents over time, and each of those incidents could include different sets of issues. So this use case is not covered by the existing “merge issues” function.