Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[testbed/mockdatasenders/mockdatadogagentexporter] Broken load test #33577

Closed
crobert-1 opened this issue Jun 14, 2024 · 7 comments · Fixed by #33860
Closed

[testbed/mockdatasenders/mockdatadogagentexporter] Broken load test #33577

crobert-1 opened this issue Jun 14, 2024 · 7 comments · Fixed by #33860
Assignees
Labels

Comments

@crobert-1
Copy link
Member

Component(s)

testbed, testbed/mockdatasenders/mockdatadogagentexporter

Describe the issue you're reporting

The load tests GitHub action is currently broken with the following error from this line:

Error when evaluating 'strategy' for job 'loadtest'. .github/workflows/load-tests.yml (Line: 64, Col: 15): Unexpected value ''

Upon further investigation, the failing variable expansion is a result of the .github/workflows/scripts/setup_e2e_tests.sh failing the splitloadtests step of the dependent action:

Run ./.github/workflows/scripts/setup_e2e_tests.sh
main module (github.com/open-telemetry/opentelemetry-collector-contrib/testbed) does not contain package github.com/open-telemetry/opentelemetry-collector-contrib/testbed/mockdatasenders/mockdatadogagentexporter
/bin/bash: line 1: -v: command not found
make: *** [Makefile:5: list-tests] Error 1
@crobert-1 crobert-1 added needs triage New item requiring triage ci-cd CI, CD, testing, build issues labels Jun 14, 2024
Copy link
Contributor

Pinging code owners:

  • testbed/mockdatasenders/mockdatadogagentexporter: @boostchicken
  • testbed: @open-telemetry/collector-approvers

See Adding Labels via Comments if you do not have permissions to add labels yourself.

@codeboten
Copy link
Contributor

This has been broken at least since February :(

@mowies
Copy link
Member

mowies commented Jul 2, 2024

@codeboten if that's fine with you, I'd like to look into this :)
maybe you could re-run one of the latest load test run from main branch for me with debug logging enabled? then i could investigate the job matrix that's not liked by github actions

@mowies
Copy link
Member

mowies commented Jul 2, 2024

Actually, I managed to debug and investigate this on my personal fork, with success it seems 👍🏼

@mowies
Copy link
Member

mowies commented Jul 2, 2024

I filed a fix PR for this @codeboten @crobert-1 :)

The array of test names was not parsed correctly in the setup_e2e_tests.sh script.
I added a more detailed description in the PR.

codeboten pushed a commit that referenced this issue Jul 2, 2024
**Description:**
This PR fixes the load test pipeline which was broken for a while.
The array of tests was not iterated over correctly in
https://github.com/open-telemetry/opentelemetry-collector-contrib/blob/main/.github/workflows/scripts/setup_e2e_tests.sh
and the first iteration of the loop resulted in the whole array being
added to the matrix instead of just the first entry of the array.

I introduced a different way to read in the array of tests at the start
of the setup bash script, which works now.
[source for the solution](https://stackoverflow.com/a/10586169)

**Link to tracking Issue:** Fixes #33577

**Testing:** Tested successfully on my personal fork. This is a [link to
a load test
run](https://github.com/mowies/opentelemetry-collector-contrib/actions/runs/9761599340)
where the matrix is correctly parsed by github actions.

Signed-off-by: Moritz Wiesinger <moritz.wiesinger@dynatrace.com>
@mowies
Copy link
Member

mowies commented Jul 3, 2024

Now, that the load tests are running again, it seems that they have some other issues as well. Pipeline runs are still red...

Taking this run for example: https://github.com/open-telemetry/opentelemetry-collector-contrib/actions/runs/9769350633

The load tests are successfully executed but something is still broken with the result upload.

@mowies
Copy link
Member

mowies commented Jul 4, 2024

@codeboten @crobert-1 I filed a follow-up PR that fixes the rest of the load test pipeline. Although I noticed that some tests are broken..
#33885

mx-psi pushed a commit that referenced this issue Jul 4, 2024
**Description:** <Describe what has changed.>
<!--Ex. Fixing a bug - Describe the bug and how this fixes the issue.
Ex. Adding a feature - Explain what this achieves.-->
This PR fixes more issues with the load test pipeline. Specifically, it
adjusts some things around artifact up/download which is needed as a
result of the breaking changes in GitHub Actions artifacts listed
[here](https://github.com/actions/upload-artifact#breaking-changes:~:text=Uploading%20to%20the%20same%20named,you%20will%20encounter%20an%20error.).
Also, it removes the issue generation step which seems far outdated as
it's failing with missing CircleCI env variables.

Also, I noticed that a few of the load tests fail consistently. To test.
the pipeline, I had to comment them out. Not sure, what to do with
them...

**Link to tracking Issue:**
Part of
#33577
Follow up to
#33860

**Testing:** Testing done on my personal fork. Link to working pipeline
run:
[Green run with some broken tests commented out, but pipeline
working](https://github.com/mowies/opentelemetry-collector-contrib/actions/runs/9789652148)
[Red run with broken
tests](https://github.com/mowies/opentelemetry-collector-contrib/actions/runs/9790395589)

**Documentation:** none added

---------

Signed-off-by: Moritz Wiesinger <moritz.wiesinger@dynatrace.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging a pull request may close this issue.

3 participants