chore: include suite, platform and tags in test reports name#2273
chore: include suite, platform and tags in test reports name#2273mdelapenya merged 2 commits intoelastic:mainfrom
Conversation
| def envContext = [] | ||
| envContext.add("PROVIDER=${provider}") | ||
| envContext.add("SUITE=${suite}") | ||
| envContext.add("REPORT_PREFIX=${suite}_${platform}_${tags}") |
There was a problem hiding this comment.
Same as the parallel stage name
.ci/scripts/functional-test.sh
Outdated
| BEAT_VERSION=${4:-"${BASE_VERSION}"} | ||
| ELASTIC_AGENT_VERSION=${5:-"${BASE_VERSION}"} | ||
| GOARCH=${GOARCH:-"amd64"} | ||
| REPORT_PREFIX="${SUITE}_${GOARCH}_${TAGS}" |
There was a problem hiding this comment.
What do you think to allow overriding the value only if REPORT_PREFIX is not set? IIUC,
https://github.com/elastic/e2e-testing/pull/2273/files#diff-ef8b2f5e4d21d81c0eba4754dea3a3dfc2c67d26284f195165b650cf4aff7beaR564 uses .ci/scripts/functional-test.sh with only one parameter, so REPORT_PREFIX="${SUITE}_${GOARCH}_${TAGS}" = REPORT_PREFIX=${tags}_ amd64_" but REPORT_PREFIX is already in place as it's created on the fly in the jenkinsfile
There was a problem hiding this comment.
That could be fine too. My only concern is that the script that is run on CI is generated on the fly in the Ansible code, and this is different to the committed script. I'd try to align them as much as possible
* chore: include suite, platform and tags in test reports name * chore: read existing value first (cherry picked from commit e516dc9)
* chore: include suite, platform and tags in test reports name * chore: read existing value first (cherry picked from commit e516dc9)
* chore: include suite, platform and tags in test reports name * chore: read existing value first (cherry picked from commit e516dc9)
* main: (72 commits) fix: always quote variables in shell scripts (elastic#2284) chore: bring back system integration scenarios (elastic#2233) chore: do not notify build green for fleet nightly builds (elastic#2277) chore: include suite, platform and tags in test reports name (elastic#2273) bump stack version 8.2.0-dcff22d7 (elastic#2270) bump stack version 8.2.0-4509f321 (elastic#2265) chore: bump Godog to v0.12.4 (elastic#2245) bump stack version 8.2.0-5cc993c1 (elastic#2254) bump stack version 8.2.0-82b803a1 (elastic#2250) fix: never process elasticsearch CI artifacts (elastic#2246) bump stack version 8.2.0-ff67d7b8 (elastic#2242) ci: periodic builds for the last two minor versions (elastic#2241) bump stack version 8.2.0-9bac538c (elastic#2240) fix: add end-of-file line in templates (elastic#2237) chore: add Christos to SSH users bump stack version 8.2.0-fee3b8d2 (elastic#2234) Separate ES workload from Agent spec and make it req for the scenarios (elastic#2203) bump stack version 8.2.0-d02c907a (elastic#2228) bump stack version 8.2.0-63265ec9 (elastic#2225) fix: wrong link (elastic#2220) ...
What does this PR do?
It adds a prefix to the test report names so that it's easier to look them up in the Jenkins UI. In the past, the report name was created calling
unamein the remote VM, now it uses the metadata coming from the platform descriptorsWhy is it important?
Consistency about file names, and adding discoverability when reading reports in the Jenkins UI
Checklist
make unit-test), and they are passing locallymake noticein the proper directory)Related issues