Check-in — 2026-08-09 session delta
TL;DR
Control-plane runner gha-mgmt / homelab-mgmt-01 is live on
runner-management. TP-Link EP10 path is fully retired (hardware dead; owner
scrubbed HA/UDM leftovers). Prox 24h mem soak still waits on Monday finalize
ntfy before shrink/retire. Stage 7 OneUptime pilot remains on ≥1 week soak.
Phase status
| Item |
Status |
| Prox mem soak (#262/#266) |
Sampling live; finalize Mon 2026-08-10 11:15 PDT |
| Alert noise / Stage 7 pilot / Komodo ntfy |
Live (see prox soak entry) |
gha-mgmt + RUNNER_FLEET_MAINTENANCE_RUNS_ON |
✅ Online 172.29.240.15; var set |
| Aqara / TP-Link 7R leftovers |
✅ Closed — EP10 deprecated; owner HA/UDM cleanup done |
| Dep bump cryptography + pymdown (#268) |
Merged |
| Hyper-V-host Ansible path (#267) |
Merged |
What shipped (this afternoon)
- Owner TODOs: create
gha-mgmt, set fleet maintenance runs-on, close Aqara
HA + EP10 (#269,
#270).
- Live device map: EP10 under Retired / removed, not “still needs move”.
- Fleet smoke:
Runner fleet operations on homelab-mgmt-01 via
RUNNER_FLEET_MAINTENANCE_RUNS_ON.
- Fleet-wide
inventory failed: GET …/ci-toolkit/actions/runners → 404
(read App likely missing Actions access on that repo)
(run).
- Repo-scoped
health scope=homelab green on homelab-mgmt-01
(run).
Known gaps / drift
- Soak journal “merge #262” rows are stale — already merged; wait for finalize
report only.
- June journal snapshots still mention deferred TP-Link (period truth; leave).
- Second
runner-management host still recommended before fleet-wide applied
maintenance (self-maintenance guard).
- Fleet-wide observe/inventory needs GitHub App install (or remove
ci-toolkit-* from fleet.yml)
before scope=fleet stays green.
What remains
| Who |
Work |
Priority |
| Owner |
Reply to Mon finalize ntfy: approve graylog / haos-influx / both / wait |
P0 |
| Agent |
After approve: shrink/retire with explicit destroy confirms |
P0 |
| Owner |
Stage 7 pilot soak ≥1 week → cutover |
P1 |
| Owner/Agent |
Fix fleet-wide observe: App access to ci-toolkit or drop instances from fleet.yml |
P1 |
| Agent |
Optional second mgmt runner before disruptive fleet-wide apply |
P2 |
flowchart LR
mgmt[gha-mgmt homelab-mgmt-01]
var[RUNNER_FLEET_MAINTENANCE_RUNS_ON]
fleet[Fleet inventory/health/maintain]
soak[Prox soak finalize Mon]
decide[Owner approve shrink]
var --> mgmt
mgmt --> fleet
soak --> decide