Is there a way to specify a pipeline stage or group to not halt further testing? I have a bunch of tests that take a few minutes each that are run parallel, I also have a few groups of these tests that in turn run serially and these all run after a stage that takes a while. I’d like to gather as many errors as possible before marking a job as failed, without actually ignoring any failures.
My pipeline looks something like:
lint1:
group: linters
lint2:
group: linters
build-x86:
group: build
build-aarch64:
group: build
vmtest-x86:
group: vmtests
vmtest-aarch64:
group: vmtests
upload-test-assets:
provision-baremetal0-ubuntu14:
group: baremetal0
provision-baremetal0-ubuntu16:
group: baremetal0
provision-baremetal1-ubuntu14:
group: baremetal1
....
check:
commands:
- ls *.failed && cat *.failed && exit 1
Where builds and uploads take a while, not to mention provisions of previous groups, and I’d like the job to continue and give me a report at the end. I’m handling it now by echo a message onto the shared workspace in each provision stage, but it would be nice if I could see the red X vs a green check for the stage.