plane(1.5): placement probe -- Substrate places by any free worker, not by fit

ops/ax/probe-placement.sh scales ousast-pool to one worker, applies probe-pool
(1 x 2 CPU / 3Gi, the already-built ateom-gvisor image, no ko build), runs one
model-free repo-facts Task requesting 2560Mi (limit 3Gi) and one requesting
256Mi, records `kubectl ate get workers -o json` after each resume, and removes
everything it created. Live on kind: both actors landed on the 1536Mi
ousast-pool worker (status.allocated.actors=1 there, 0 on probe-pool) ->
placement: any. design.md section 4 records the outcome and the fallback group
5 must take (an engine atespace with engine-pool as its only pool if Substrate
scopes pools by atespace; else a plain Kubernetes Job on k3s). Evidence under
benchmarks/measurements/2026-10-02-k8s-01-profile/probe/.

Gates (tree identical to 1.4 in src/ and tests/; ops/ gains the probe script,
which the literal scan covers): pytest 2010 passed, 42 skipped, exit 0; ruff
check 0; ruff format 0; mypy 0.

Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_01EmM1j46qgKMh3zqD1QMwEo
This commit is contained in:
Marius Ciepluch
2026-10-02 20:49:03 +02:00
co-authored by Claude Fable 5.1
parent 8099b29deb
commit 86ec237483
9 changed files with 387 additions and 1 deletions
+1 -1
View File
@@ -42,7 +42,7 @@ go-ahead** with a ceiling; steps that need a hand on a cluster that does not exi
`runner-image`; `generate.repin_templates` accepts that file (runner key only until task 4.3).
- Evidence: `ops/ax/up.sh` reruns idempotently on kind; `ops/ax/smoke-run.sh` passes.
- _Requirements: 1.1, 2.4_
- [ ] 1.5 Placement probe (open question 2, decided before the engine work)
- [x] 1.5 Placement probe (open question 2, decided before the engine work) -- result: `placement: any`
- `ops/ax/probe-placement.sh`: applies a second WorkerPool `probe-pool` (1 replica, 3 Gi) from the design's
engine template, then one model-free `repo-facts` Task with `resources.requests.memory: 2560Mi`, and records
with `kubectl ate get workers` which worker took the actor; then the same Task with 256Mi. Removed after.