ci(workflow): use sized Warp runners
Withdrawn to preserve free public CI
The cost audit does not justify this migration: standard GitHub-hosted runners are free for this public repository, while Warp adds paid usage. No measured time benefit establishes a worthwhile tradeoff. Closing this PR preserves the existing GitHub workflows and the migration branch/history. No runner-group access change is needed.
Estimated adopted change: $0 added runner cost and 0 change in execution time. The original proposal and estimates below are retained as history, not a recommendation to merge or expand runner access.
Summary
Move 2 compatible job definitions in 2 workflows to WarpBuild. Use 2 vCPU / 8 GB for small Linux jobs and 4 vCPU / 16 GB where builds, tests, or memory requirements justify it. Preserve OS versions, architecture, triggers, permissions, commands, and existing Warp allocations.
Runner choices
| Workflow / job | Runner | Reason |
|---|---|---|
validate.yml / validate | warp-ubuntu-latest-x64-2x | 2 vCPU / 8 GB for short serial checks and orchestration; these jobs do not need the 4 CPUs supplied by standard public GitHub runners. |
release-please.yml / release-please | warp-ubuntu-latest-x64-2x | 2 vCPU / 8 GB for short serial checks and orchestration; these jobs do not need the 4 CPUs supplied by standard public GitHub runners. |
Estimated time and cost
These are planning scenarios, not measured Warp speedups. The conservative case assumes unchanged execution time; sensitivity cases assume 20% and 40% shorter execution. A 10-minute job would save 0, 2, or 4 minutes respectively. Network waits, cloud deployments, and API polling may see little or no speedup. Runner-minute savings are not the same as pipeline wall-clock savings when jobs run concurrently.
This repository is public: standard GitHub-hosted runners currently cost $0. Warp introduces runner spend; there is no direct runner-bill saving. The potential benefit is execution time and consistent runner infrastructure.
| Per 1,000 current runner minutes, before per-job rounding | Current | Warp: same runtime | Warp: 20% shorter | Warp: 40% shorter |
|---|---|---|---|---|
ubuntu-latest → warp-ubuntu-latest-x64-2x | $0.000 | $4.000 | $3.200 | $2.400 |
Historical sample: 6 successful changed-job executions, drawn from up to three recent successful runs per workflow within the first 100 repository runs returned for 7 September–7 October 2026. This is a selected sample, not complete monthly usage. Some workflows have changed since those historical runs, so it is an indicative baseline rather than a controlled benchmark.
Replaying that sample gives 1.30 runner minutes and $0.000 current list-price equivalent. Warp would cost $0.024 at the same runtime, $0.024 at 20% shorter, or $0.024 at 40% shorter. Corresponding modeled time savings are 0 / 0.26 / 0.52 minutes. Each job is rounded up separately to a whole billable minute.
| Sample scenario | Modeled runner-bill saving (negative = added spend) |
|---|---|
| Same runtime | $-0.024 |
| 20% shorter runtime | $-0.024 |
| 40% shorter runtime | $-0.024 |
Sample sources: run 36693486531, run 36691307760, run 36691196265.
| Job | Successful observations | Median execution |
|---|---|---|
validate.yml / validate | 3 | 0.20 min |
release-please.yml / release-please | 3 | 0.22 min |
Rates checked 7 October 2026: Warp runner catalog, GitHub Actions billing, Warp billing increment. Excludes discounts, credits, tax, storage, and external service charges. No monthly dollar claim is made without complete usage and allowance data.
Validation
- Parsed every changed workflow and compared its structure with the base: only
runs-onor the documented reusable-workflowwith.runs-oninput changes. - Checked runner labels, OS/architecture compatibility, and preserved matrix identities and job conditions.
- Independently reviewed the migration and corrected serial jobs that did not need 4 CPUs.
- Runtime CI and automated reviewer feedback will be checked on this PR. Release/deployment-only paths still need their normal post-merge execution; no release or deployment has been dispatched to test this change.
Before merging, confirm the Warp GitHub App can access this repository and the checks start on the requested labels. Public repositories additionally need access to the default runner group. If jobs remain queued, enable repository access before merging; do not bypass required checks.
Summary by cubic
Moves validate and release-please jobs to smaller Warp runners (warp-ubuntu-latest-x64-2x, 2 vCPU / 8 GB) to reduce runner spend while keeping OS, architecture, triggers, and permissions unchanged.
This repository is public, so GitHub-hosted runners are free today; Warp introduces runner cost. The benefit is potential execution-time savings from consistent Warp infrastructure. Estimated modeled impact for a 1,000-minute baseline is $4.00 at the same runtime, or $3.20/$2.40 at 20%/40% shorter runtimes. Historical sampling shows these jobs are very short (median 0.20–0.22 min), so modeled added spend is about $0.024 per 1,000 runner-minutes.
Validation
- Only
runs-onvalues changed; workflow structure, matrix identities, and job conditions are preserved. - Confirm Warp GitHub App access to this repository and that checks start on the requested labels before merging.