I agree, this seems like a good generic drone feature. I got hit by the same slack outage ([Duplicate] Non fatal pipeline steps) but could see this also happening with other steps in our pipeline (updating jira issues, etc) that are not indicative of a “failed” build.
If other’s agree, I’d be happy to take a stab at implementing something and PRing it with a bit of guidance.
This capability is supported in the new yaml and runtime packages by adding allow_failure: true to the step. These new packages will be introduced in 0.9