Add filter when checking out OpenSearch-Dashboards to speed up CI checks

[cwperks]

Description

Adds filter to the actions/checkout@v2 step of install-dashboards to minimize the size of the repo that is cloned. The cypress folder is over 600MB in size and not needed for the security integ tests.

Category

Test fix

Issues Resolved

#1611

Check List

• New functionality includes testing
• New functionality has been documented
• Commits are signed per the DCO using --signoff

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.

[codecov]

Codecov Report

Merging #1613 (f91773f) into main (d3b3f84) will increase coverage by 1.75%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             main    #1613      +/-   ##
==========================================
+ Coverage   66.18%   67.93%   +1.75%     
==========================================
  Files          93       93              
  Lines        2339     2339              
  Branches      317      317              
==========================================
+ Hits         1548     1589      +41     
  Misses        722      722              
+ Partials       69       28      -41     

see 17 files with indirect coverage changes

[cwperks]

The integ test failures are definitely due to high watermark.

Filesystem      Size  Used Avail Use% Mounted on
/dev/root        84G   74G  9.5G  89% /

Where is most of the space being consumed?

[cwperks]

4 / 10 of the windows runs failed on the first run, but succeeded on the 2nd. Its fairly stable with this fix, but still susceptible to failure. I'm filing an issue now around stability and speed for the CI on windows.

[cwperks]

@peternied issue on this repo: #1620

I am also going to file an issue in OSD core to see if its possible to only install prod dependencies.

Something like NODE_ENVIRONMENT=production yarn osd bootstrap

[peternied]

4 / 10 of the windows runs failed on the first run, but succeeded on the 2nd. Its fairly stable with this fix, but still susceptible to failure. I'm filing an issue now around stability and speed for the CI on windows.

Wait, what was the failure rate before? I know you've been investing time in this - but I'd rather we make the minimal change with low downside rather that potential create more issues.

Let me know if you've got before and after numbers - I am also happy to fold these changes into the overall CI improvements work and let you move onto other topics.

[cwperks]

@peternied The failure rate before this change was 100% on Windows.

Edit: I just kicked off 10 iterations for each platform before this change here: https://github.com/cwperks/security-dashboards-plugin/actions/runs/6577075426

[cwperks]

@peternied Please see the link above. Only 2 / 10 passed before the change to free up space on the Windows runner.

[peternied]

On its own the change to filter check out of the Dashboards repository - happy to merge.

Manually deleting/uninstall - I'm not comfortable without a 100% pass rate. It's hard to measure/maintain as a testament to the work you've done - I don't want us chasing future maintenance issues in the space if it isn't the full picture.

If we don't have the Windows platform operating reliably, it sounds like we should just disable it and then separately see how to get it reliable.

[cwperks]

If we don't have the Windows platform operating reliably, it sounds like we should just disable it and then separately see how to get it reliable.

I would be supportive of that. This change would make the pass rate go from <20% to >70% on Windows so it would provide benefit in the interim while #1620 is worked on.

btw if any uninstall fails, the workflow continues as normal and would effectively work as it does today where it would run with all of the pre-installed software.

[DarshitChanpura]

@peternied @cwperks I've update the workflows to disable disk threshold check and have also reverted changes related to deleting pre-installed software on Windows runners

[cwperks]

Looks good to me!

[peternied]

@DarshitChanpura Thanks for your help on this. I've created the code instability tracing issue [1] as well as a number of observations about test pass rate. Lets try to keep working against a known state - while you are focused on this maybe you could update and track where we are at until we can get back to reliable.

• [1] Stabilize CI for dashboards repo #1624