coulomb/railiance-cluster
4
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Publish pending unscheduled pods in cluster observations.
All checks were successful
CI Smoke / host-smoke (push) Successful in 0s
Details
CI Smoke / container-smoke (push) Successful in 2s
Details

Browse source 
RAIL-BS-WP-0014: residual millicores count scheduled pods only.
state_hub_preflight_observation matches the STATE-WP-0091 fixture shape.

Assistant: grok
Assistant-Session: 01a09dc1-b21e-77e1-919e-fcad2f82b267
This commit is contained in:
codex 2026-09-14 10:34:27 +02:00
parent 551a898e4d
commit b36ac28b49
4 changed files with 121 additions and 12 deletions
Download patch file Download diff file Expand all files Collapse all files

12
workplans/RAIL-BS-WP-0014-state-hub-surge-headroom.md
View file

@ -4,7 +4,7 @@ type: workplan
title: "Keep node remaining CPU honest for State Hub surge and pending demand"
domain: financials
repo: railiance-cluster
status: proposed
status: finished
owner: codex
topic_slug: railiance
origin: residual
@ -12,6 +12,7 @@ origin_ref: STATE-WP-0091
created: "2026-09-14"
updated: "2026-09-14"
related: [RCLUSTER-WP-0014, STATE-WP-0091, CUST-WP-0071]
state_hub_workstream_id: "d89863aa-dc89-5a9d-be00-4f558a16d70b"
---
Residual from STATE-WP-0091. State Hub now refuses promotion when remaining
@ -24,11 +25,18 @@ not capacity admission for factory or other zero-request workloads.
```task
id: RAIL-BS-WP-0014-T01
status: todo
status: done
priority: high
state_hub_task_id: "d90c384e-8a99-550f-b8b8-6314c073f2dc"
```
Extend `tools/observe_cluster_resources.py` / `make cluster-observe` so the
published observation names pending unscheduled pods and their requests.
State Hub's preflight already consumes that shape. Do not treat residual
millicores as a scheduling guarantee.
**Done 2026-09-14.** `build_observation` names pending unscheduled pods, uses
effective pod requests (init vs containers), and emits
`state_hub_preflight_observation` in the STATE-WP-0091 fixture shape.
Residual CPU counts scheduled pods only. `not_a_scheduling_guarantee` is
explicit. Unit test covers a 50m pending pod that does not inflate remaining.