This feature would account for temporary outages by persisting check data between runs, and counting the amount of consecutive failures each link has.
Ideally, the feature should allow users to configure the amount of times a link may consecutively fail and be forgiven. However, locking N to 2 here would simplify implementation as we wouldn't have to keep track of how many times a link has failed, only that it did in the last run.
To do either, two capabilities have to be added:
-
The ability to save link check data and load it between runs, Artifacts are a good candidate.
-
The ability to have a link fail without causing the whole test to report as failed.
While this ticket is for the issue in the title, it's worth noting Artifacts capability may also provide a more stable reporting interface than Checks and Action CLI output.
This feature would account for temporary outages by persisting check data between runs, and counting the amount of consecutive failures each link has.
Ideally, the feature should allow users to configure the amount of times a link may consecutively fail and be forgiven. However, locking N to 2 here would simplify implementation as we wouldn't have to keep track of how many times a link has failed, only that it did in the last run.
To do either, two capabilities have to be added:
The ability to save link check data and load it between runs, Artifacts are a good candidate.
The ability to have a link fail without causing the whole test to report as failed.
While this ticket is for the issue in the title, it's worth noting Artifacts capability may also provide a more stable reporting interface than Checks and Action CLI output.