Compare commits

...

10 commits

Author SHA1 Message Date
df6ed8a520 Add Forgejo CI smoke and image workflow (tier 3 T04)
All checks were successful
CI Smoke / host-smoke (push) Successful in 0s
CI Smoke / container-smoke (push) Successful in 1s
Build and Publish Container Image / build-and-push (push) Successful in 1m11s
2026-07-06 15:54:32 +02:00
5371eaf139 chore(consistency): sync task status from DB [auto]
Updated by fix-consistency on 2026-07-03:
  - update .custodian-brief.md for core-hub
2026-07-03 18:36:27 +02:00
c77706c94c chore(consistency): sync task status from DB [auto]
Updated by fix-consistency on 2026-07-03:
  - update .custodian-brief.md for core-hub
2026-07-03 11:25:59 +02:00
173e8d9cfc CORE-WP-0005-T04 done: hub.coulomb.social cut over to Core Hub (verified, rollback retained)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-03 11:25:01 +02:00
defe4a6af8 Production deployment profile + cutover runbook for hub.coulomb.social (CORE-WP-0005-T04)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-03 01:41:38 +02:00
a33191ab3a CORE-WP-0005-T04 prerequisites resolved: seed full Inter-Hub vocabulary; keep /hubs protected (ops-hub gate expects 401)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-03 01:39:45 +02:00
0a23c9a9b4 CORE-WP-0005-T04: record cutover prerequisites from T03 (catalog vocab + auth posture)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 19:25:24 +02:00
a43b9fe640 chore(consistency): sync task status from DB [auto]
Updated by fix-consistency on 2026-07-02:
  - update .custodian-brief.md for core-hub
2026-07-02 19:24:39 +02:00
1f1288ad94 CORE-WP-0005-T03 done: dual-run compat deltas; found ops-* widget-type catalog gap for cutover
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 19:24:09 +02:00
a26abe8334 chore(consistency): sync task status from DB [auto]
Updated by fix-consistency on 2026-07-02:
  - update .custodian-brief.md for core-hub
2026-07-02 19:21:38 +02:00
10 changed files with 461 additions and 11 deletions

View file

@ -2,7 +2,7 @@
# Custodian Brief — core-hub
**Domain:** infotech
**Last synced:** 2026-07-02 16:13 UTC
**Last synced:** 2026-07-03 16:36 UTC
**State Hub:** http://127.0.0.1:8000 *(adjust if running on a remote machine)*
## Active Workstreams
@ -16,14 +16,6 @@ Progress: 0/4 done | workstream_id: `a8d66822-e435-4b1e-ad81-37a298d1795e`
- ! Update Cross-Repo References `ed7f8e01`
- · Retire Probe Repo `d3a494a3`
### Migrate Inter-Hub data and cut over production
Progress: 1/4 done | workstream_id: `a36efc08-5906-40c7-b1de-f3a553ffea02`
**Open tasks:**
- ! Run Staging Import `e0ea0928`
- ! Dual-Run Smokes `9166afe8`
- ! Production Cutover `fa68858a`
---
## MCP Orientation (when available)

View file

@ -0,0 +1,29 @@
# Canonical CI smoke template (tier 1 routing drill).
# Copy to: .forgejo/workflows/ci-smoke.yaml in consumer repos.
name: CI Smoke
on:
push:
branches:
- main
workflow_dispatch:
jobs:
host-smoke:
runs-on: self-hosted
steps:
- name: Routing probe (host runner)
run: |
set -eu
echo "repository=${GITHUB_REPOSITORY:-unknown}"
echo "sha=${GITHUB_SHA:-unknown}"
echo "runner=${RUNNER_NAME:-unknown}"
uname -a
container-smoke:
runs-on: ubuntu-latest
steps:
- name: Routing probe (container label)
run: |
set -eu
echo "container-smoke ok for ${GITHUB_REPOSITORY:-unknown}"

View file

@ -0,0 +1,47 @@
name: Build and Publish Container Image
on:
push:
branches:
- main
paths:
- ".forgejo/workflows/image.yaml"
- "Dockerfile"
- "src/**"
- "migrations/**"
- "scripts/**"
- "pyproject.toml"
- "uv.lock"
- "alembic.ini"
workflow_dispatch:
env:
REGISTRY: forgejo.coulomb.social
IMAGE_NAME: coulomb/core-hub
DOCKER_HOST: tcp://127.0.0.1:2375
jobs:
build-and-push:
runs-on: container-build
steps:
- name: Build and push image
env:
REGISTRY_USER: ${{ secrets.REGISTRY_USER }}
REGISTRY_TOKEN: ${{ secrets.REGISTRY_TOKEN }}
run: |
set -eu
REF="${GITHUB_SHA:-main}"
SHORT="${REF:0:7}"
mkdir -p buildctx "${HOME}/bin"
wget -qO /tmp/repo.tar.gz \
"https://forgejo.coulomb.social/${GITHUB_REPOSITORY}/archive/${SHORT}.tar.gz"
tar xzf /tmp/repo.tar.gz -C buildctx --strip-components=1
wget -qO- https://download.docker.com/linux/static/stable/x86_64/docker-27.3.1.tgz \
| tar xz --strip-components=1 -C "${HOME}/bin" docker/docker
export PATH="${HOME}/bin:${PATH}"
echo "${REGISTRY_TOKEN}" | docker login "${REGISTRY}" -u "${REGISTRY_USER}" --password-stdin
IMAGE="${REGISTRY}/${IMAGE_NAME}"
docker build -t "${IMAGE}:latest" -t "${IMAGE}:main-${SHORT}" buildctx
docker push "${IMAGE}:latest"
docker push "${IMAGE}:main-${SHORT}"
echo "pushed ${IMAGE}:latest and ${IMAGE}:main-${SHORT}"

View file

@ -0,0 +1,9 @@
apiVersion: v1
kind: Namespace
metadata:
name: core-hub
labels:
app.kubernetes.io/name: core-hub
app.kubernetes.io/part-of: core-hub
environment: production
railiance.io/postgres-client: apps-pg

View file

@ -0,0 +1,113 @@
apiVersion: v1
kind: ConfigMap
metadata:
name: core-hub-prod-runtime
namespace: core-hub
labels:
app.kubernetes.io/name: core-hub
app.kubernetes.io/part-of: core-hub
environment: production
data:
CORE_HUB_ENV: production
CORE_HUB_AUTO_CREATE_TABLES: "0"
---
apiVersion: batch/v1
kind: Job
metadata:
name: core-hub-prod-migrate
namespace: core-hub
labels:
app.kubernetes.io/name: core-hub
app.kubernetes.io/component: migration
app.kubernetes.io/part-of: core-hub
environment: production
spec:
backoffLimit: 1
template:
metadata:
labels:
app.kubernetes.io/name: core-hub
app.kubernetes.io/component: migration
app.kubernetes.io/part-of: core-hub
environment: production
spec:
restartPolicy: Never
containers:
- name: migrate
image: gitea.coulomb.social/coulomb/core-hub:production-placeholder
imagePullPolicy: IfNotPresent
command: ["alembic", "upgrade", "head"]
envFrom:
- configMapRef:
name: core-hub-prod-runtime
- secretRef:
name: core-hub-prod-env
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: core-hub-api
namespace: core-hub
labels:
app.kubernetes.io/name: core-hub
app.kubernetes.io/component: api
app.kubernetes.io/part-of: core-hub
environment: production
spec:
replicas: 1
selector:
matchLabels:
app.kubernetes.io/name: core-hub
app.kubernetes.io/component: api
template:
metadata:
labels:
app.kubernetes.io/name: core-hub
app.kubernetes.io/component: api
app.kubernetes.io/part-of: core-hub
environment: production
spec:
containers:
- name: api
image: gitea.coulomb.social/coulomb/core-hub:production-placeholder
imagePullPolicy: IfNotPresent
ports:
- name: http
containerPort: 8010
envFrom:
- configMapRef:
name: core-hub-prod-runtime
- secretRef:
name: core-hub-prod-env
readinessProbe:
httpGet:
path: /readyz
port: http
initialDelaySeconds: 5
periodSeconds: 10
livenessProbe:
httpGet:
path: /healthz
port: http
initialDelaySeconds: 15
periodSeconds: 20
---
apiVersion: v1
kind: Service
metadata:
name: core-hub-api
namespace: core-hub
labels:
app.kubernetes.io/name: core-hub
app.kubernetes.io/component: api
app.kubernetes.io/part-of: core-hub
environment: production
spec:
type: ClusterIP
selector:
app.kubernetes.io/name: core-hub
app.kubernetes.io/component: api
ports:
- name: http
port: 8010
targetPort: http

View file

@ -0,0 +1,30 @@
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: core-hub
namespace: core-hub
annotations:
cert-manager.io/cluster-issuer: letsencrypt-prod
traefik.ingress.kubernetes.io/router.entrypoints: websecure
traefik.ingress.kubernetes.io/router.tls: "true"
labels:
app.kubernetes.io/name: core-hub
app.kubernetes.io/part-of: core-hub
environment: production
spec:
ingressClassName: traefik
rules:
- host: hub.coulomb.social
http:
paths:
- backend:
service:
name: core-hub-api
port:
number: 8010
path: /
pathType: Prefix
tls:
- hosts:
- hub.coulomb.social
secretName: core-hub-tls

View file

@ -0,0 +1,47 @@
# Core Hub production deployment (hub.coulomb.social cutover)
`30-ingress.yaml` is the cutover flip — do NOT apply it while the inter-hub
ingress still claims `hub.coulomb.social`.
## Sequence
```bash
# 1. Image (operator: registry push)
IMAGE_REPOSITORY=gitea.coulomb.social/coulomb/core-hub IMAGE_TAG=<sha> make container-build
docker push gitea.coulomb.social/coulomb/core-hub:<sha>
# 2. Prod DB + secret (operator: shared-DB DDL)
# role core_hub / db core_hub on apps-pg; secret core-hub-prod-env with
# CORE_HUB_DATABASE_URL (postgresql+asyncpg://...apps-pg-rw.databases...:5432/core_hub)
# and CORE_HUB_API_TOKEN (generated)
# 3. Deploy (no traffic yet)
kubectl apply -f k8s/railiance-production/00-namespace.yaml
kubectl apply -f k8s/railiance-production/20-runtime.yaml
kubectl -n core-hub set image job/core-hub-prod-migrate migrate=gitea.coulomb.social/coulomb/core-hub:<sha>
kubectl -n core-hub set image deploy/core-hub-api api=gitea.coulomb.social/coulomb/core-hub:<sha>
kubectl -n core-hub wait --for=condition=complete job/core-hub-prod-migrate --timeout=180s
kubectl -n core-hub rollout status deploy/core-hub-api --timeout=180s
# 4. Final data import (fresh Inter-Hub export at cutover time)
kubectl exec -i -n databases net-kingdom-pg-1 -- psql -U postgres -d interhub -tA \
< scripts/migration/inter_hub_export.sql > final-bundle.json
# validate, then import in-cluster (see scripts/migration/README.md)
# 5. THE FLIP (operator approval point)
kubectl -n inter-hub delete ingress inter-hub
kubectl apply -f k8s/railiance-production/30-ingress.yaml
# cert-manager issues core-hub-tls for hub.coulomb.social (~1 min)
# 6. Verify
curl -s https://hub.coulomb.social/healthz # 200
curl -s https://hub.coulomb.social/api/v2/widget-types | head -c 200
curl -s -o /dev/null -w '%{http_code}' https://hub.coulomb.social/api/v2/hubs # 401 = hardened contract
## Rollback (any time; Inter-Hub deployment stays running)
kubectl delete -f k8s/railiance-production/30-ingress.yaml
kubectl -n inter-hub apply -f <saved inter-hub ingress> # or helm upgrade inter-hub
```
Keep the Inter-Hub deployment/service running untouched until stabilization
passes (CORE-WP-0005-T04); only the ingress moves.

View file

@ -0,0 +1,39 @@
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
annotations:
cert-manager.io/cluster-issuer: letsencrypt-prod
meta.helm.sh/release-name: inter-hub
meta.helm.sh/release-namespace: inter-hub
traefik.ingress.kubernetes.io/router.entrypoints: websecure
traefik.ingress.kubernetes.io/router.tls: "true"
creationTimestamp: "2026-06-14T12:49:01Z"
generation: 1
labels:
app: inter-hub
app.kubernetes.io/managed-by: Helm
name: inter-hub
namespace: inter-hub
resourceVersion: "14479609"
uid: 2f1d1557-76be-445c-b77b-e03e7b0c3935
spec:
ingressClassName: traefik
rules:
- host: hub.coulomb.social
http:
paths:
- backend:
service:
name: inter-hub
port:
number: 8000
path: /
pathType: Prefix
tls:
- hosts:
- hub.coulomb.social
secretName: inter-hub-tls
status:
loadBalancer:
ingress:
- ip: 92.205.130.254

View file

@ -35,12 +35,69 @@ WIDGET_TYPES = [
CatalogItem(slug="status-summary", name="Status Summary"),
CatalogItem(slug="workplan-board", name="Workplan Board"),
CatalogItem(slug="event-stream", name="Event Stream"),
# Inter-Hub compatibility vocabulary (CORE-WP-0005-T04 prerequisite 1):
# the full legacy widget_type_registry so migrated widgets stay
# referentially complete and legacy catalog consumers see a superset.
CatalogItem(slug="action", name="Action Control", description="Button, link, or trigger widget"),
CatalogItem(slug="chart", name="Chart", description="Data visualisation chart widget"),
CatalogItem(slug="chat", name="Chat Region", description="Conversational interaction region"),
CatalogItem(slug="diff", name="Diff / Review", description="Code diff or change review element"),
CatalogItem(slug="form", name="Form", description="Data entry form widget"),
CatalogItem(slug="ops-backup-set", name="ops-backup-set"),
CatalogItem(slug="ops-cluster", name="ops-cluster"),
CatalogItem(slug="ops-endpoint", name="ops-endpoint"),
CatalogItem(slug="ops-environment", name="ops-environment"),
CatalogItem(slug="ops-host", name="ops-host"),
CatalogItem(slug="ops-incident", name="ops-incident"),
CatalogItem(slug="ops-migration-wave", name="ops-migration-wave"),
CatalogItem(slug="ops-readiness-gate", name="ops-readiness-gate"),
CatalogItem(slug="ops-release", name="ops-release"),
CatalogItem(slug="ops-risk", name="ops-risk"),
CatalogItem(slug="ops-runbook", name="ops-runbook"),
CatalogItem(slug="ops-secret-set", name="ops-secret-set"),
CatalogItem(slug="ops-service", name="ops-service"),
CatalogItem(slug="ops-service-catalog", name="ops-service-catalog"),
CatalogItem(slug="panel", name="Status Panel", description="Summary or status information panel"),
CatalogItem(slug="recommendation", name="Recommendation", description="AI or system recommendation block"),
CatalogItem(slug="table", name="Table", description="Tabular data display widget"),
CatalogItem(slug="workflow-step", name="Workflow Step", description="Single step in a multi-step workflow"),
]
EVENT_TYPES = [
CatalogItem(slug="interaction.event", name="Interaction Event"),
CatalogItem(slug="workplan.progress", name="Workplan Progress"),
CatalogItem(slug="evidence.recorded", name="Evidence Recorded"),
# Inter-Hub compatibility vocabulary (see WIDGET_TYPES note).
CatalogItem(slug="abandoned", name="Abandoned"),
CatalogItem(slug="accepted_recommendation", name="Accepted Recommendation"),
CatalogItem(slug="blocked_by_policy", name="Blocked by Policy"),
CatalogItem(slug="clicked", name="Clicked"),
CatalogItem(slug="commented", name="Commented"),
CatalogItem(slug="escalated", name="Escalated"),
CatalogItem(slug="failed", name="Failed"),
CatalogItem(slug="flagged_confusing", name="Flagged Confusing"),
CatalogItem(slug="flagged_helpful", name="Flagged Helpful"),
CatalogItem(slug="focused", name="Focused"),
CatalogItem(slug="ops-backup-verified", name="ops-backup-verified"),
CatalogItem(slug="ops-drift-detected", name="ops-drift-detected"),
CatalogItem(slug="ops-endpoint-verified", name="ops-endpoint-verified"),
CatalogItem(slug="ops-health-checked", name="ops-health-checked"),
CatalogItem(slug="ops-inventory-registered", name="ops-inventory-registered"),
CatalogItem(slug="ops-inventory-updated", name="ops-inventory-updated"),
CatalogItem(slug="ops-migration-gate-failed", name="ops-migration-gate-failed"),
CatalogItem(slug="ops-migration-gate-passed", name="ops-migration-gate-passed"),
CatalogItem(slug="ops-readiness-gate-updated", name="ops-readiness-gate-updated"),
CatalogItem(slug="ops-release-observed", name="ops-release-observed"),
CatalogItem(slug="ops-restore-tested", name="ops-restore-tested"),
CatalogItem(slug="ops-risk-accepted", name="ops-risk-accepted"),
CatalogItem(slug="ops-risk-raised", name="ops-risk-raised"),
CatalogItem(slug="ops-runbook-executed", name="ops-runbook-executed"),
CatalogItem(slug="ops-service-discovered", name="ops-service-discovered"),
CatalogItem(slug="rejected_recommendation", name="Rejected Recommendation"),
CatalogItem(slug="retracted", name="Retracted"),
CatalogItem(slug="retried", name="Retried"),
CatalogItem(slug="submitted", name="Submitted"),
CatalogItem(slug="viewed", name="Viewed"),
]
ANNOTATION_CATEGORIES = [

View file

@ -47,7 +47,7 @@ actual import can run.
```task
id: CORE-WP-0005-T03
status: wait
status: done
priority: high
state_hub_task_id: "9166afe8-27ef-4345-a7b1-c33f15c90673"
```
@ -60,13 +60,46 @@ Run public and protected smokes against both Inter-Hub and Core Hub. Record comp
```task
id: CORE-WP-0005-T04
status: wait
status: done
priority: high
state_hub_task_id: "fa68858a-b18e-4f09-a477-6b11e54099d4"
```
With operator approval, switch production traffic such as `hub.coulomb.social` to Core Hub. Keep rollback available until stabilization passes.
**Prerequisites (blocking, found in T03 dual-run 2026-07-02):**
1. **Migrate/seed the ops-* type vocabularies.** Migrated Inter-Hub widgets
reference 7 widget types (`ops-endpoint`, `ops-environment`, `ops-host`,
`ops-migration-wave`, `ops-readiness-gate`, `ops-service`,
`ops-service-catalog`) plus their event types, none of which exist in Core
Hub's seeded registries. Either extend the `core-hub.migration.v1` bundle
with `widgetTypeRegistry` / `eventTypeRegistry` collections, or seed the
ops-* vocabularies into Core Hub before cutover, or migrated widgets are
referentially incomplete.
2. **Auth-posture change is a consumer break.** `/api/v2/hubs` is public on
Inter-Hub but protected on Core Hub — any anonymous consumer of the hub
list must gain a token before cutover.
3. Operator approval + rollback plan (existing gate).
**Prerequisite resolutions 2026-07-03:**
1. **Catalog gap — resolved by seeding.** Core Hub's catalogs are static
in-code seeds (`seeds.py`), not DB registries, so extending the migration
bundle schema was the wrong shape. Seeded the complete Inter-Hub
vocabulary instead (23 widget types incl. all 7 orphaned `ops-*` types,
30 event types) as an explicit Inter-Hub-compatibility block; existing
Core Hub seeds retained. 22 tests pass. Migrated widgets are
referentially complete once the new image ships.
2. **Auth posture — resolved by keeping Core Hub protected.** Consumer
audit found only two `/api/v2/hubs` status consumers:
`ops-hub/interhub_gate_probe.py` **asserts 401** (it encodes the desired
hardened contract and currently fails against Inter-Hub production,
which serves 200), and `railiance-forge/tools/gitea-runner-status.sh`
is informational. No anonymous consumer depends on a 200. Core Hub's
protected listing is therefore an intentional hardening that satisfies
the ops-hub gate at cutover; no code change made.
## Progress 2026-07-02 — staging import completed
@ -84,3 +117,57 @@ second run proved idempotency (created=0, updated=28, no duplicates).
Server-side confirmed: the imported `ops-hub` hub and its 15 widgets are visible
via `/api/v2/*`. T02 done; the staging DB now holds real Inter-Hub data for the
T03 dual-run smokes.
## Progress 2026-07-02 — dual-run smokes and compatibility deltas
Ran unauthenticated + authenticated probes against both live surfaces via
ops-bridge tunnels (Inter-Hub `inter-hub-coulombcore`:18020, Core Hub staging
`core-hub-staging-coulombcore`:18010).
Status parity table (unauthenticated GET):
| Endpoint | Inter-Hub | Core Hub | Note |
| --- | --- | --- | --- |
| `/healthz`, `/readyz` | 404 | 200 | Core Hub adds liveness/readiness (new capability, not a regression). |
| `/api/v2/hubs` | 200 | 401 | **Auth delta**: hub listing is public on Inter-Hub, protected on Core Hub. Anonymous consumers must authenticate against Core Hub. |
| `/api/v2/hub-registry` | 401 | 401 | Parity. |
| `/api/v2/widget-types`, `event-types`, `annotation-categories`, `policy-scopes` | 200 | 200 | Public-catalog status parity. |
| `/api/v2/widgets`, `interaction-events` | 401 | 401 | Parity. |
Content/shape deltas:
- Catalog record shape differs: Core Hub uses `{name, slug, description}`;
Inter-Hub uses `{name, label, description, ownerHubId, status}`. Consumers
parsing catalog records need updates for the `slug`-based shape.
- **Cutover blocker found**: the migrated Inter-Hub widgets reference 7
widget-type names (`ops-endpoint`, `ops-environment`, `ops-host`,
`ops-migration-wave`, `ops-readiness-gate`, `ops-service`,
`ops-service-catalog`) that are **absent** from Core Hub's seeded
widget-type registry (which has only `Event Stream`, `Status Summary`,
`Workplan Board`). The `core-hub.migration.v1` bundle schema (COLLECTIONS)
carries widgets but not the `widget_type_registry` / `event_type_registry`,
so migrated widgets land referentially incomplete. **Production cutover
(T04) must first migrate or seed the ops-* widget-type and event-type
vocabularies into Core Hub.** Filed as the top T04 prerequisite.
T03 done: compatibility surface is characterized; the auth-hardening delta and
the missing-catalog-vocabulary blocker are the two items the cutover decision
must account for.
## Completion 2026-07-03 — production cutover executed
Operator-approved and operator-executed flip: `hub.coulomb.social` ingress
moved from Inter-Hub to Core Hub (`core-hub` namespace, image `a33191a`).
Verified post-flip: cert-manager issued `core-hub-tls`; `/healthz` and
`/readyz` 200; public catalogs 200 serving 26 widget types (all formerly
orphaned `ops-*` types present); `/api/v2/hubs` 401 — the hardened contract
the ops-hub gate probe asserts. Post-flip idempotent reconcile import: 0
created / 28 updated / 0 errors (no Inter-Hub writes were missed).
Rollback remains one command (`30-ingress.yaml` delete +
`rollback-inter-hub-ingress.yaml` apply) and the Inter-Hub deployment stays
running untouched through the stabilization window. `CORE-WP-0007` Haskell
retirement stays gated until stabilization passes and the operator approves
retiring the fallback.