Skip to content

docs(clusters): document k3s-rabbit/oc-ampere e2e coverage and fix stale oc-ampere description - #1798

Merged
dark-vex merged 3 commits into
mainfrom
docs/k3s-rabbit-oc-ampere-e2e
Aug 7, 2026
Merged

docs(clusters): document k3s-rabbit/oc-ampere e2e coverage and fix stale oc-ampere description#1798
dark-vex merged 3 commits into
mainfrom
docs/k3s-rabbit-oc-ampere-e2e

Conversation

@dark-vex

@dark-vex dark-vex commented Aug 6, 2026

Copy link
Copy Markdown
Owner

Summary

Companion to #1796 (validate-k3s-rabbit.yml) and #1797 (validate-oc-ampere.yml).

  • clusters/CLAUDE.md: adds both new workflows to the CI table, fixes the stale "Teleport agent only" oc-ampere cluster-table description (it has 4 real apps) and the same staleness in the Slack-alerting "not extended" rationale, and documents the shared EXCLUDED_APPS-first fail-closed per-app loop plus why no concurrency: guard is needed across the 4 cluster e2e workflows — confirmed both self-hosted runners (gen8-runner/Lugano, psp-runner/BGY) are distinct physical machines that each process one job at a time, so GitHub's own scheduling already prevents same-machine collisions.
  • Both new clusters' root kustomization.yaml gain the same "this resources: list is parsed by ..." header comment kubenuc/k8s-vms-daniele already carry.
  • clusters/k3s-rabbit/CLAUDE.md: documents the "zero apps deploy by design" framing for its e2e workflow.
  • clusters/oc-ampere/CLAUDE.md: new file (didn't exist before) — documents the ARM64-production/x86_64-CI mismatch, the real 4-app inventory, the no-1Password gap, and the flux-operator pre-FluxInstance safety finding from ci(validate-oc-ampere): add e2e cluster validation workflow #1797.

Test plan

🤖 Generated with Claude Code

…ale oc-ampere description

Companion to the e2e workflow PRs adding validate-k3s-rabbit.yml and
validate-oc-ampere.yml:

- clusters/CLAUDE.md: add both workflows to the CI table, fix the
  stale "Teleport agent only" oc-ampere cluster-table description (it
  has 4 real apps), fix the same staleness in the Slack-alerting
  "not extended" rationale, and document the EXCLUDED_APPS-first
  fail-closed per-app loop shared across all four cluster e2e
  workflows plus why no concurrency guard is needed (both self-hosted
  runners are confirmed distinct physical machines).
- Both new clusters' root kustomization.yaml gain the same
  "this resources: list is parsed by ..." header comment kubenuc/
  k8s-vms-daniele already carry, so a new top-level resource doesn't
  silently lose e2e coverage. oc-ampere's also documents why
  flux-instance.yaml must stay skip-listed there specifically.
- clusters/k3s-rabbit/CLAUDE.md: document that its e2e workflow
  deploys zero apps by design (3/3 real apps excluded) and what the
  run still proves.
- clusters/oc-ampere/CLAUDE.md: new file (didn't exist before).
  Documents the ARM64-production/x86_64-CI mismatch, the real app
  inventory, the no-1Password gap, and the flux-operator
  pre-FluxInstance safety finding from the e2e PR.

Signed-off-by: Daniele De Lorenzi <2905124+dark-vex@users.noreply.github.com>
@dark-vex
dark-vex merged commit f2712d5 into main Aug 7, 2026
24 checks passed
@dark-vex
dark-vex deleted the docs/k3s-rabbit-oc-ampere-e2e branch August 7, 2026 14:16
dark-vex added a commit that referenced this pull request Aug 7, 2026
docs(clusters): document k3s-rabbit/oc-ampere e2e coverage and fix stale oc-ampere description
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant