chore: Improve image build speed

[34fathombelow]

Signed-off-by: Justin Marquis [email protected]

Disabled cache, it is configured to only cache the export layers that are in the final build stages. Not taking advantage of the cache on disk for the second run of the kubectl-argo-rollouts image This will reduce build times by half. I recommend a second PR to cleanup the existing cache steps from the docker-publish.yml as they are not needed anymore.

I have tested these builds locally. I'm expecting the complete Workflow should be complete in approximately 20-25 minutes on GitHub runners.

The flowchart below shows how the Dockerfile is processed using cross-compilation. Orange nodes are x86_64 and opaque orange nodes are arm64.

flowchart TB

id1([builder]) --> id2([argo-rollouts-ui]) --> id3([argo-rollouts-build]) --> id4([kubectl-argo-rollouts]) --> id5([x86_64 Final_image])
id3([argo-rollouts-build]) --> id6([kubectl-argo-rollouts]) --> id7([arm64 Final_Image])


style id6 fill:#5C74C6,stroke:#333,stroke-width:2px,color:#ffffff
classDef x86 fill:#EF7B4D,stroke:#EF7B4D,stroke-width:2px,color:#ffffff
classDef arm64 fill:#FCE5DB,stroke:#EF7B4D,stroke-width:2px,color:#EF7B4D
class id1,id2,id3,id4,id5 x86
class id6,id7,id8 arm64

Loading

Checklist:

• Either (a) I've created an enhancement proposal and discussed it with the community, (b) this is a bug fix, or (c) this is a chore.
• The title of the PR is (a) conventional, (b) states what changed, and (c) suffixes the related issues number. E.g. "fix(controller): Updates such and such. Fixes #1234".
• I've signed my commits with DCO
• I have written unit and/or e2e tests for my change. PRs without these are unlikely to be merged.
• My builds are green. Try syncing with master if they are not.
• My organization is added to USERS.md.

[codecov]

Codecov Report

Merging #1919 (1209b0a) into master (08cf10e) will increase coverage by 0.03%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #1919      +/-   ##
==========================================
+ Coverage   82.36%   82.40%   +0.03%     
==========================================
  Files         119      119              
  Lines       16872    16910      +38     
==========================================
+ Hits        13897    13934      +37     
- Misses       2284     2285       +1     
  Partials      691      691              

Impacted Files	Coverage Δ
controller/metrics/metrics.go	100.00% <0.00%> (ø)
controller/metrics/prommetrics.go	100.00% <0.00%> (ø)
utils/record/record.go	80.55% <0.00%> (+2.83%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 08cf10e...1209b0a. Read the comment docs.

[sonarcloud]

Kudos, SonarCloud Quality Gate passed!   

0 Bugs
0 Vulnerabilities
0 Security Hotspots
0 Code Smells

No Coverage information
No Duplication information

[harikrongali]

is this intentional to remove cache?

[34fathombelow]

Yes, the cache is only caching the exported layers in the final build stage. Which provides no real benefit. This could be changed to mode=max to export all interminate layers but GitHub has a limit of 10GB across a single repo. Also the time to prossess approximately 20gb of cache and compress it really outweighs the benefits. A full cache will also not prossess RUN commands in the dockerfile.

[34fathombelow]

Also this cache is being accessed during pull requests. It can create conflicts and slow the processes down.

[harikrongali]

got it..thanks

[harikrongali]

thanks for the PR

[34fathombelow]

@harikrongali @jessesuen Thank you for the review and merge. Image build times are much better, but the plugin-image should have been built from the runners cache of the controller-image. I will look into this and have a solution by Monday. We should be able to cut off an additional 20+ minutes.