mirror of
https://github.com/norandom/OpenUltraSAST.git
synced 2026-10-16 11:49:28 +00:00
Marius CiepluchandClaude Opus 5.5
34da916030
`ousast plane harvest` (plane/harvest.py) writes two Runs over a label snapshot:
verify a/b + agree on the verify-family units (both sides of each labelled pair as
an opaque files-only Workspace, candidates outside the hunt's root, no declared
sites; v1 and dev-php pins with repo-facts first) and model roles on the candidate
files (pair excerpts grouped, one Task per population pin). Budgets sum under the
ceiling by construction. New roles task template; the generator re-pins it.
Spent by attribution: verify $2.54 (1961 calls, 150 units, 182 hunts per pass),
roles $1.40 (1034 calls, 799 files, 0 unclassified chunks); DeepSeek balance
19.63 -> 18.23. Engine on the pair corpus (benchmarks/learn/engine_pairs.py,
frozen e6ed785, one 3g container) is PARTIAL: 384 of 686 sides at commit, still
running; unfinished sides stay missing. benchmarks/learn/build_features.py built
1712 static+plane records into the memory store; counts-only record under
benchmarks/measurements/2026-10-01-decision-engine-harvest/.
source/roles parts are `none`, not `failed`, for a language without a declaration
pattern (C). ops/ax/README: Substrate golden-actor disk leak and the inline
Workspace limit (~20 KB).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01EmM1j46qgKMh3zqD1QMwEo
75 lines
5.4 KiB
Markdown
75 lines
5.4 KiB
Markdown
# ax on this host
|
|
|
|
The service plane (`.kiro/specs/ai-service-plane/`) runs every task on google/ax over Agent Substrate in a
|
|
single-node kind cluster. ax is the only executor; `ousast plane run` submits, resumes, starts and collects.
|
|
|
|
## Bring-up
|
|
|
|
ops/ax/up.sh # idempotent: kind + registry, Substrate, ax, gVisor worker pool, egress gateway,
|
|
# receiver Service, runner image (digest pinned), ax CLI, smoke Task
|
|
ousast plane doctor # kind, Substrate, ax controller, runner image
|
|
ops/ax/smoke-run.sh # end to end: repo-facts on ax, artifacts back on the host, attribution table
|
|
ops/ax/down.sh # deletes the cluster and its registry
|
|
|
|
Tools live in `~/go/bin` (kind, ax, ko, kubectl-ate) and `~/.local/bin` (kubectl); sources and rendered files
|
|
in `~/.cache/ousast/ax-src/` (the runner digest pin is `runner-image` there).
|
|
|
|
## Measured footprint (2026-09-29, idle, two workers)
|
|
|
|
| Item | Value |
|
|
|---|---|
|
|
| kind node container memory | 1.7 GiB (27% of the 7.7 GiB host) |
|
|
| registry container memory | 34 MiB |
|
|
| node volume (images, containerd) | 4.9 GB on disk |
|
|
| worker pool | 2 gVisor workers, 1 CPU / 1.5 GiB limit each |
|
|
|
|
Disk is the binding constraint: the Substrate install and `ko` builds need several GB of Go caches. Clear them
|
|
afterwards (`go clean -cache -modcache`); rerunning the full Substrate install on a near-full disk fails in
|
|
`ko resolve`.
|
|
|
|
## What this host taught (each one cost a failed live run)
|
|
|
|
- Substrate installs no WorkerPool ("no free workers") -> `workerpool.yaml.tmpl`.
|
|
- Actor images must be pinned by digest; ax's default runner image on gcr.io needs credentials.
|
|
- A release `ax` CLI skews from the server; build it from the deployed checkout.
|
|
- ax's API rejects unknown fields; metadata is `name` + `atespace`; Git entries have no `commit`.
|
|
- `ax apply` leaves a Task Suspended until `ax resume`; the first resume may time out while Substrate builds the
|
|
template's golden snapshot. ax has no Completed phase: completion is our artifact delivery.
|
|
- Substrate boots each template once as a golden actor with the task's env, so the runner waits for a start
|
|
request sent through `atenet-router`; the provider key travels only in that request.
|
|
- Actor egress goes through the `atenet-egress` gateway, deny by default, one EgressPolicy per actor, hostnames
|
|
only; the gateway dials the address the actor connected to, so the receiver is a Service dialled by ClusterIP
|
|
with its name as Host. The Envoy gateway needs a Rust build; the agentgateway variant needs none.
|
|
- Apply a Task after the Workspaces it binds; ax copies them at create time.
|
|
- Every Task is its own actor template with a golden actor (~24 MB under `/var/lib/ate/actors` on the node). Deleting
|
|
them failed while the snapshot bucket ax's templates name (`dberkov-gke-dev3`, from `AX_SNAPSHOTS_BUCKET`) did not
|
|
exist in rustfs: golden actors piled up in `DELETING` and filled the disk during a 600-task harvest (2026-10-01).
|
|
Create that bucket in rustfs (an aws-cli pod with the `rustfs-bucket-init` env); even then a deleted actor's
|
|
directory stays on the node, so long Runs need a janitor that removes directories no live actor owns.
|
|
- A Workspace's `files` reach the actor inline in one environment variable: above ~20 KB of content the template
|
|
fails with "actor template not found" (an 86 KB and a 31 KB excerpt did; 7 KB ran).
|
|
|
|
## Moving to a separate Kubernetes cluster (planned)
|
|
|
|
The maintainer will deploy the plane to its own Kubernetes environment. The manifests under `plane/` move as they
|
|
are; these parts assume this laptop and must change first (checked 2026-09-30):
|
|
|
|
| Assumption | Where | Needed in a real cluster |
|
|
|---|---|---|
|
|
| Artifact receiver runs on the developer host, reached through the `ousast-receiver` Service with an EndpointSlice to the kind gateway `172.19.0.1:18090` | `egress.py`, `receiver-service.yaml.tmpl` | the receiver as an in-cluster Deployment (or object storage, e.g. Substrate's S3-compatible store), with the reconciler reading from it |
|
|
| Registry `localhost:5001`, rewritten by Substrate for kind | `up.sh`, `doctor.py:45`, runner digest pin | a registry the workers can pull from; keep digest pins (Substrate rejects tags) |
|
|
| kubectl context `kind-$KIND_CLUSTER_NAME` | `doctor.py:39`, `egress.py`, `router.py` | one configurable context (for example `OUSAST_KUBE_CONTEXT`) |
|
|
| ax's snapshot bucket: `AX_SNAPSHOTS_BUCKET` in ax's `deploy/ax-server.yaml` points at the ax authors' GCS bucket | ax deploy manifest | your own bucket, set before deploying ax |
|
|
| Egress gateway applied by hand (agentgateway variant, no Rust build) | this README | the Substrate-installed gateway; per-task EgressPolicies work unchanged |
|
|
| Worker pool of 2 x 1 CPU / 1.5 GiB for the 7 GB host | `workerpool.yaml.tmpl` | sized to the cluster; `--workers` to match |
|
|
| Plane memory store and results under `~/ousast-results/` on the host | reconciler, memory layer (harnessx-removal Req 6) | a persistent volume or bucket shared by the reconciler |
|
|
|
|
The provider key already travels only in the start request through `atenet-router`, which works the same through
|
|
a port-forward to any cluster.
|
|
|
|
## What a larger ax deployment needs
|
|
|
|
A Kubernetes cluster with Agent Substrate (and its egress gateway), the ax control plane, a registry the workers
|
|
can pull from, the runner image pinned by digest, the receiver reachable as a Service, and the provider host
|
|
allowed per task. The manifests under `plane/` move unchanged; only `OUSAST_ARTIFACT_*` and the context change.
|