[Feature Request] Support multiple open incidents + customizable incident name
Is your feature request related to a problem? Please describe.
Yes. Currently, the Incident Manager in OpenMetadata only displays the most recent incident, even if multiple test cases or metrics are failing simultaneously. This limits observability and makes it hard to track multiple active problems at the same time, especially in large environments where several tables, metrics, or pipelines may raise alerts.
Describe the solution you'd like I’d like the ability to:
Display multiple open incidents in the UI, not just the latest one.
Optionally customize the name of each incident, either manually or automatically (e.g. include test case name, metric, table, etc.).
View a list or panel of all current incidents with status, timestamps, and context.
Describe alternatives you've considered A clear and concise description of any alternative solutions or features you've considered.
Additional context This would help a lot in scenarios where multiple test cases run per day with different metrics. Right now, only the last one is visible, making others disappear from the incident view.
@aimendenche-nw currently the logic is to assign the open incident to the any subsequent test failure for a specific test case up until it is resolved.
e.g. Test case fails on day 1 for test case A -> we open incident Z Test case fails on day 2 for test case A and incident Z was not resolved -> we don't open a new incident and keep incident Z as open day 3 incident Z is resolved but test case A fails agains -> we will open a new incident Y
LMK if the logic above makes sense -- happy to hear your feedback.
@aimendenche-nw please respond to @TeddyCr comments above. If we don’t hear back we would be closing this as “works as designed"
hey @TeddyCr @harshach T
hanks for the explanation! The current logic makes sense in terms of deduplication per test case, but in our use case, we’d prefer a different behavior.
We actually want a new incident to be created for each test execution, even if a previous incident for the same test case is still open. The idea is to retain a full historical record of incidents, one per failed run, to better track the progression and evolution of issues over time.
For example, we run a DAG daily that pushes test case results to OpenMetadata. If a test fails three days in a row, we’d like to see three distinct incidents, each representing the failure at a specific point in time , rather than reusing and updating the same one.
In short, we’d like an option to:
Allow multiple open incidents per test case
Create a new incident per test run, regardless of existing unresolved ones
Optionally customize incident names to include timestamp, test name, or table name...
hello @TeddyCr, as we discussed last time about this feature, here’s the document detailing the proposa. it aligns with several of the needs we mentioned. You can review and organize things as you see fit, depending on what’s most practical for you.