circleci: make implicit-medium optimism jobs explicit (12 jobs, mixed gen2 sizes)

[hdcesario-op]

Summary

Removes the last 12 jobs in this repo that declared no resource_class: field. Post-#20917 / #20927 every other job is on an explicit Gen2 class; these were the remaining stragglers running on Gen1 medium by CircleCI default. Each is set to an explicit Gen2 class right-sized to its workload, plus two machine-image pins to preserve the Gen1 OS where it would otherwise have silently upgraded jammy → noble.

This completes the optimism repo's Gen2 migration. No further resource_class work is planned here.

Why Gen2, not Gen1

Post-#20927 the whole repo is on Gen2. Mixing Gen1 in for these last 12 jobs would create two-class drift we'd just have to clean up later.

Why one PR, not two phases

The original handoff (and #20917's pattern) anticipated ~110 implicit-medium jobs, which would have justified splitting "make explicit" (Phase A) from "right-size" (Phase B). The actual scope is 12 jobs — small enough that per-job rationale fits in one audit-grade PR.

Per-job rationale

File	Job	Executor	Class	Image change	Why this class
.circleci/config.yml	prepare-continuation-config	docker	medium.gen2	—	Setup-pipeline job at the start of every CI run. Does a full monorepo checkout. Conservative: blast radius if it OOMs is every pipeline.
.circleci/continue/main.yml	ci-gate	docker	small.gen2	—	Single utils/ci-gate call. Pure API gate, deterministic workload.
.circleci/continue/main.yml	l2-chains-sync-check	docker	small.gen2	—	One bash script (check-l2-chains-sync.sh) verifying two files are in sync.
.circleci/continue/main.yml	todo-issues	machine	medium.gen2	+ image: ubuntu-2204:current	Machine executor minimum is medium. Image param c-base_image resolves to the literal string "default", so swapping class would silently switch host from jammy to noble. Pin preserves Gen1 OS.
.circleci/continue/main.yml	cannon-prestate	machine	medium.gen2	+ image: ubuntu-2204:current	Machine, build job (just reproducible-prestate), and no machine.image declared today — so the highest-risk job for the Gen1→Gen2 OS-default change. Pin first; broader right-sizing question left for a follow-up.
.circleci/continue/main.yml	cannon-stf-verify	docker	medium.gen2	—	Has setup_remote_docker, which is not supported on small.gen2, plus a cannon build step.
.circleci/continue/main.yml	op-deployer-forge-version	docker	small.gen2	—	Single just check-forge-version call.
.circleci/continue/main.yml	stale-check	machine	medium.gen2	—	Machine minimum is medium. Image already pinned to ubuntu-2204:2024.08.1 — no OS-default risk.
.circleci/continue/main.yml	close-issue	machine	medium.gen2	—	Same as stale-check: machine minimum, image already pinned.
.circleci/continue/rust-ci.yml	required-rust-ci	docker	small.gen2	—	Pure aggregator gate (utils/ci-gate only).
.circleci/continue/rust-e2e.yml	required-rust-e2e	docker	small.gen2	—	Pure aggregator gate (utils/ci-gate only).
.circleci/rust-nightly-bump.yml	bump-nightly	docker	small.gen2	—	Daily-scheduled shell + git + gh pr create. Pure version bump.

Totals: 6 → small.gen2, 6 → medium.gen2, 2 image pins.

Methodology for picking small.gen2

Bias toward determinism-of-workload, not just utilization numbers. From the #20899 / #20931 fuzz lesson: telemetry doesn't capture outlier-seed spikes for distribution-tailed workloads, so a job that averages low but is non-deterministic should stay larger than the averages suggest.

Applied here as a checklist — small.gen2 only if all of these are true:

• Job name pattern fits (*-gate, *-check, version-bump, *-aggregator, prepare/setup)
• Steps contain no go test / cargo test / forge test / build commands
• No setup_remote_docker (CircleCI does not support it on small.gen2)
• No parallelism: N (none of the 12 declare this anyway)
• Executor is docker (machine has no small tier)

Of the 12 jobs, 6 cleared every condition.

Machine-executor image-pin notes

Per the #20927 zstd lesson: machine jobs that don't explicitly pin machine.image inherit the class default (Gen1 → ubuntu-2204:current / jammy, Gen2 → ubuntu-2404:current / noble). Swapping <class> → <class>.gen2 therefore silently upgrades the host OS.

Two of the four machine jobs in this PR were exposed to that:

• cannon-prestate declared no image: at all.
• todo-issues referenced <<pipeline.parameters.c-base_image>>, which defaults to the literal string "default" — CircleCI's sentinel for "use the class default".

Both now pin image: ubuntu-2204:current to preserve the Gen1 OS. stale-check and close-issue were already on pinned ubuntu-2204:2024.08.1 and need no image change.

Validation gate

• 14 insertions, 1 deletion across 5 files (git diff --stat)
• 12 jobs that previously had no resource_class: field now have one
• 0 implicit-class jobs remain in any of the 5 CI YAML files (re-run of the enumerator)
• All 5 modified YAMLs still parse with yaml.safe_load
• No other diff content: nothing reordered, no unrelated edits

Refs

• circleci: migrate fuzz-golang to medium.gen2 (W4 pilot) #20899 (W4 pilot, reverted)
• circleci: migrate optimism Tier A jobs to Gen2 (same-size, 41 docker jobs across 3 files) #20917 (Tier A: 41 same-size docker swaps)
• circleci: complete optimism Gen2 migration (22 jobs across 3 files) #20927 (Final: 22 same-size docker + machine swaps)
• Resize Fuzz-golang job to xlarge.gen2 #20931 (fuzz-golang resize back to xlarge.gen2)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.1%. Comparing base (e7c16a1) to head (052f03b).

Additional details and impacted files

@@            Coverage Diff             @@
##           develop   #20957     +/-   ##
==========================================
- Coverage     76.3%    76.1%   -0.3%     
==========================================
  Files          185      185             
  Lines        10712    10712             
==========================================
- Hits          8178     8156     -22     
- Misses        2390     2412     +22     
  Partials       144      144             

Flag	Coverage Δ
cannon-go-tests-64	66.3% <ø> (ø)
contracts-bedrock-tests	82.0% <ø> (-0.4%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.
see 1 file with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.