Commit Graph

170 Commits

Author SHA1 Message Date
Will Anderson dda07ae269 Route /oauth2, /connect, /.well-known at neurontechnologies.ai to MCP 2026-04-27 17:51:59 -05:00
Will Anderson 03495962e0 fix(daemon): use neuron-rest cluster service instead of local JVM (--jar was circular proxy) 2026-04-27 17:26:58 -05:00
Will Anderson 3943ebc204 feat: scale marketing to 1 replica, add notification + turnstile secrets
Single replica required for in-memory pending leads Map used by SMS
double-opt-in flow. ExternalSecret adds Resend, Twilio, and Turnstile
keys to the marketing pod environment.
2026-04-27 03:31:45 -05:00
will.anderson b2940c70e0 fix(swarm): add runAsUser: 1001 to all swarm containers (#44) 2026-04-27 05:13:50 +00:00
Neuron CI e9da888512 ci(neuron-prod): update rest+license to v0.18.14 2026-04-26 17:26:16 +00:00
Neuron CI cdadd723b6 ci(neuron-prod): blue-green flip to blue@v0.18.14 (registry.neuralplatform.ai/neuron-technologies/neuron-mcp:v0.18.14) 2026-04-26 17:26:05 +00:00
Neuron CI 41f94d93be ci(daemon): deploy main neuron-daemon@0f5dbd6a 2026-04-26 17:02:49 +00:00
Neuron CI 9d7bc48ed0 ci(daemon): deploy main neuron-daemon@10cd48e0 2026-04-26 16:55:20 +00:00
Neuron CI 15482b9241 ci(neuron-prod): update rest+license to v0.18.8 2026-04-26 16:49:46 +00:00
Neuron CI bd7d5efa66 ci(neuron-prod): blue-green flip to green@v0.18.8 (registry.neuralplatform.ai/neuron-technologies/neuron-mcp:v0.18.8) 2026-04-26 16:49:02 +00:00
Neuron CI 726510d7d9 ci(daemon): deploy main neuron-daemon@5dc2d833 2026-04-26 16:34:07 +00:00
will.anderson 66bffc5a4b Merge pull request 'fix(rest): run as UID 1000 to match PVC file ownership' (#35) from feat/daemon-image-c190e6e5 into main 2026-04-26 16:07:02 +00:00
Will Anderson c6245d1170 fix(rest): run as UID 1000 to match PVC file ownership — SQLite needs write access 2026-04-26 11:06:23 -05:00
will.anderson f056806cbd Merge pull request 'fix(rest): Recreate strategy to avoid SQLite write lock on rollout' (#34) from feat/daemon-image-c190e6e5 into main 2026-04-26 16:05:24 +00:00
Will Anderson 58513835e2 fix(rest): use Recreate strategy to avoid SQLite write lock on rollout 2026-04-26 11:04:16 -05:00
Neuron CI f525437c77 ci(daemon): deploy main neuron-daemon@b2bed846 2026-04-26 16:03:29 +00:00
Will Anderson baf147543c ci(daemon): deploy main neuron-daemon@c190e6e5 2026-04-26 10:37:29 -05:00
Will Anderson 714a7c7a8d launch: deploy daemon StatefulSet, fix neuron-rest security context, add daemon Argo CD app
- Add neuron-daemon Argo CD Application (apps/neuron-daemon.yaml) pointing to
  k8s/neuron-technologies/daemon/ on main — daemon will deploy once image is built
- Add daemon k8s manifests: StatefulSet (2 replicas, imagePullPolicy: Always),
  Service (ClusterIP :7749, session affinity), ExternalSecret (Anthropic API key from Vault)
- Fix neuron-rest runAsNonRoot: remove constraint since v0.15.3 image runs as root,
  clears the CreateContainerConfigError on the 11h-old stale pod
- Add NEURON_DAEMON_WS_URL env var to marketing deployment for chat UI connectivity
- Add neuron.neurontechnologies.ai Cloudflare DNS record and tunnel ingress rule (Terraform)
2026-04-26 10:13:48 -05:00
Will Anderson 655ab38558 Deploy neuron-mcp:a837369 — VBD marketplace architecture 2026-04-26 08:51:07 -05:00
Will Anderson 6041ac3b25 Deploy neuron-mcp:514db58 — STATELESS, built from current main 2026-04-26 08:33:03 -05:00
Will Anderson 623a6828e8 Deploy neuron-mcp:05d978f — STATELESS from main branch 2026-04-26 08:24:55 -05:00
Will Anderson d804d71040 Remove sticky session from MCP IngressRoute — STATELESS is sessionless 2026-04-26 08:21:16 -05:00
Will Anderson 5f1c9adaa6 Deploy neuron-mcp:1b26e3e — STATELESS HTTP transport
Removes STREAMABLE session state. No sticky sessions needed.
Works across all pods, survives restarts, scales cleanly.
2026-04-26 08:18:07 -05:00
Will Anderson 08ef13c40d Add sticky session affinity to MCP IngressRoute
STREAMABLE transport holds session state in-memory per-pod. With HPA
scaling to 2+ replicas and round-robin routing, requests hit different
pods and fail with "Session not found". Pin each client to one pod via
a secure HttpOnly cookie.
2026-04-26 08:07:23 -05:00
Will Anderson 20397e9831 Restore neuron-mcp:97bd19ba (rollback to v0.15.3 was incompatible) 2026-04-26 04:28:19 -05:00
Will Anderson 5fa8cf45d5 Rollback neuron-mcp to v0.15.3 to restore MCP connectivity 2026-04-26 04:20:57 -05:00
Will Anderson 83b8d0768e Add Traefik IngressRoute for neuron.neurontechnologies.ai MCP endpoint 2026-04-26 03:52:24 -05:00
Will Anderson 6ac4faabc5 Deploy neuron-mcp:97bd19ba — Postgres marketplace datasource 2026-04-26 03:27:46 -05:00
Will Anderson e18d953ff9 Wire marketplace Postgres datasource for neuron-mcp
Add MARKETPLACE_DB_URL/USER to ConfigMap and MARKETPLACE_DB_PASSWORD
ExternalSecret (sourced from secret/legion-db in Vault). Remove the
SQLite subPath volume mount and fix-data-ownership initContainer from
the blue deployment — marketplace storage is now in Postgres.
2026-04-26 02:33:26 -05:00
Will Anderson 952e781550 Pre-create neuron-marketplace.db in initContainer for subPath mount
subPath mounts require the source file to exist on the PVC before the
main container starts. Add 'touch /data/neuron-marketplace.db' to the
initContainer so the subPath mount can bind the file at /app/neuron-marketplace.db.
2026-04-26 02:06:02 -05:00
Will Anderson 702faca691 Mount neuron-marketplace.db from PVC via subPath at /app/neuron-marketplace.db
App hard-codes relative path 'neuron-marketplace.db' from working dir /app.
Mount the PVC file at /app/neuron-marketplace.db via subPath so the app can
write to it as UID 1000. PVC persists the file across pod restarts.
Also keep NEURON_MARKETPLACE_DB_PATH in ConfigMap for future app versions.
2026-04-26 02:03:09 -05:00
Will Anderson a6b0dc9211 Force pod restart to pick up NEURON_MARKETPLACE_DB_PATH configmap change
Add config-hash annotation to template to cycle the ReplicaSet and
ensure pods restart with the updated ConfigMap env var.
2026-04-26 01:52:19 -05:00
Will Anderson ddaab28638 Fix marketplace DB: revert workingDir, add NEURON_MARKETPLACE_DB_PATH
workingDir: /data broke the JVM (app.jar is a relative path in /app).
Add NEURON_MARKETPLACE_DB_PATH=/data/neuron-marketplace.db env var
following the same pattern as NEURON_DB_PATH for the core DB.
2026-04-26 01:49:48 -05:00
Will Anderson 26ed92ca38 Set workingDir: /data so relative DB paths resolve to PVC
neuron-marketplace.db was opening relative to /app (image root,
not writable by UID 1000). Set workingDir to /data so all relative
file opens land on the writable PVC volume.
2026-04-26 01:47:02 -05:00
Will Anderson 754cd72df5 Fix SQLite READONLY: chown /data to UID 1000 via initContainer
PVC was written as root before runAsUser: 1000 was added; SQLite
refused writes with SQLITE_READONLY. initContainer runs as root to
chown the volume, then the app runs as UID 1000 as required.
2026-04-26 01:43:37 -05:00
Will Anderson 01245a6278 Fix Neuron MCP: runAsUser + move to neuron.neurontechnologies.ai
Pod was crash-looping: image has no USER directive so kubelet rejected it
with runAsNonRoot. Add runAsUser: 1000 at both pod and container level.

MCP paths moved from neurontechnologies.ai (now a GCP Cloud Run A record,
unreachable on Legion) to private subdomain neuron.neurontechnologies.ai.
Cloudflare tunnel and DNS CNAME updated out-of-band.
2026-04-26 01:35:20 -05:00
Will Anderson 74dd054a36 feat(swarm): add swarm self-improvement loop infrastructure
Adds complete k8s manifests, ArgoCD app, and Terraform namespace
resources for the Neuron swarm self-improvement loop system.

Each variant (alpha, beta, gamma) gets its own isolated namespace,
PVC, MCP/REST deployments, ExternalSecrets from Vault, RBAC for CI,
and a SQLite clone Job template for session startup.
2026-04-26 00:58:43 -05:00
Will Anderson 93358505fc Harden prod: security, autoscaling, observability, BuildKit CI
Security:
- Drop ALL capabilities, enforce non-root, RuntimeDefault seccomp on
  neuron-mcp, neuron-rest, neuron-marketing pods
- Add startup probes (150s window for JVM) so liveness doesn't fire early
- Replace docker-sock hostPath with BuildKit rootless TCP endpoint
  (moby/buildkit:v0.19.0-rootless) — removes node root access from CI
- Document full ESO AppRole migration path in cluster-secret-store.yaml

Autoscaling & availability:
- HPAs on mcp (1–6), rest (1–4), marketing (2–8) at 65–70% CPU
- PodDisruptionBudgets (minAvailable: 1) on all three services
- NetworkPolicy: default-deny-all in neuron-prod, explicit allow rules
  for Traefik ingress, intra-namespace, and egress to DNS/platform/vault

Observability:
- ServiceMonitors for mcp, rest, marketing (cross-namespace enabled in
  kube-prometheus-stack with serviceMonitorSelectorNilUsesHelmValues:false)
- PrometheusRules: high error rate, high latency, crash loops, replica
  shortage, Postgres down/connections, backup failure, backup staleness

Chart version pinning:
- kube-prometheus-stack, loki, tempo, redis, alloy, postgres — all pinned
  to major-version ranges to block silent breaking upgrades

Backup hardening:
- restic:latest → restic:0.17.3 (deterministic image)
- Weekly backup-verify CronJob: restores latest snapshot and validates
  SQL dump structure (≥5 CREATE TABLE, pg_dump header check)

ArgoCD:
- neuron-prod AppProject: scopes deploys to neuron-prod + platform ns,
  blacklists ClusterRole/ClusterRoleBinding/Namespace creation,
  automated sync window 2–6am UTC, manual always allowed
2026-04-25 22:54:18 -05:00
Will Anderson 7eeff54a11 Wire GCS backup to neuron-db-backup-prod bucket (neuron-494301)
Bucket created, SA key stored in Vault at secret/gcs.
CronJob ExternalSecret updated to pull from secret/gcs.
Hourly restic backup now runs to both R2 and GCS.
2026-04-25 17:51:57 -05:00
Will Anderson 8d97bbd802 Merge branch 'main' of https://git.neuralplatform.ai/will/infrastructure 2026-04-25 17:50:32 -05:00
Will Anderson a37deca724 Add GCS backup bucket + dual-destination hourly backup (R2 + GCS)
Provision Google Cloud Storage bucket for neuron prod DB backups via Terraform.
Create dedicated backup service account with objectAdmin on the bucket.
Update neuron-prod backup CronJob to run restic against both R2 and GCS hourly —
R2 as primary, GCS as secondary, independent credentials and repositories.
2026-04-25 15:23:51 -05:00
Neuron CI 8de866a8b9 ci(neuron-prod): update rest+license to v0.15.3 2026-04-25 20:06:27 +00:00
Neuron CI df85598df2 ci(neuron-prod): blue-green flip to blue@v0.15.3 (registry.neuralplatform.ai/neuron-technologies/neuron-mcp:v0.15.3) 2026-04-25 20:06:23 +00:00
Neuron CI 950f00586f ci(neuron-prod): update rest+license to v0.15.2 2026-04-25 19:53:53 +00:00
Neuron CI 35e346d94c ci(neuron-prod): blue-green flip to green@v0.15.2 (registry.neuralplatform.ai/neuron-technologies/neuron-mcp:v0.15.2) 2026-04-25 19:53:50 +00:00
Neuron CI 8e5320c7b3 ci(neuron-prod): update rest+license to v0.15.1 2026-04-25 19:39:31 +00:00
Neuron CI 9b057411db ci(neuron-prod): blue-green flip to blue@v0.15.1 (registry.neuralplatform.ai/neuron-technologies/neuron-mcp:v0.15.1) 2026-04-25 19:39:29 +00:00
Neuron CI 14f0663e28 ci(neuron-prod): update rest+license to v0.15.0 2026-04-25 19:20:40 +00:00
Neuron CI 8b56eb5290 ci(neuron-prod): blue-green flip to green@v0.15.0 (registry.neuralplatform.ai/neuron-technologies/neuron-mcp:v0.15.0) 2026-04-25 19:20:35 +00:00
Neuron CI 39050bfde4 ci(neuron-prod): update rest+license to v0.14.12 2026-04-25 10:18:30 +00:00