Enable MicroVM Benchmarking - #691
Conversation
Benjamin Elder (BenTheElder)
left a comment
There was a problem hiding this comment.
generally lgtm, I think, one confusing nit
| # 2. hack/install-microvm-deps.sh --install (asset build/stage + cluster-wide | ||
| # microvm-default SandboxConfig) | ||
| # 3. Apply the counter-microvm demo manifest (namespace + WorkerPool + | ||
| # ActorTemplate; resolves to microvm-default via sandboxClass=microvm). |
There was a problem hiding this comment.
🤖 nit 🟢 – This describes the mechanism the PR deliberately rejected. The config is named microvm, not microvm-default, and sandboxconfig-microvm.yaml.tmpl says it "is deliberately NOT marked default:true" so a dirty teardown cannot silently bind new pools — binding is by explicit sandboxConfigName, which is what the demo manifest and deploy.sh both do. "Resolves to microvm-default via sandboxClass=microvm" reads as default-by-class resolution, so a reader tracing why a WorkerPool picks this config would draw the wrong conclusion.
The stale name also appears at line 24 here and in three orchestrator.py docstrings (lines 288, 294, 441).
Currently we are unable to easily automate the benchmarking of micro VMs.
This PR: