Commit Graph

14 Commits

Author SHA1 Message Date
Will Anderson e5c9fd1287 Add Anthropic API key secret and env var to Soma Cloud Run
Provisions soma-anthropic-api-key in Secret Manager and injects it as
ANTHROPIC_API_KEY into the Cloud Run service. Also sets ANTHROPIC_MODEL
to claude-sonnet-4-5. Soma's chat handler now uses this to route the
neuron model to Anthropic instead of spinning up GPU VMs.
2026-04-28 21:29:27 -05:00
Will Anderson c3a90ac484 Add SOMA_KEY_SECRET to Secret Manager and wire it into Cloud Run
Creates soma-key-secret in GCP Secret Manager for HMAC key signing, and
adds SOMA_KEY_SECRET + SOMA_BASE_DOMAIN env vars to the soma-prod-us Cloud
Run service so stateless API key validation works in production.
2026-04-28 11:39:33 -05:00
Will Anderson 1a8e039ecf Rename inference endpoint to neuron.neurontechnologies.ai, add wildcard DNS for customer orgs 2026-04-28 11:15:11 -05:00
Will Anderson 9e2646701c Grant soma compute permissions for GCE inference nodes
- Add roles/compute.instanceAdmin.v1 for soma SA (provision/terminate GCE VMs)
- Add roles/iam.serviceAccountUser for soma SA (assign SA to inference VMs)
- Add firewall rule: allow TCP 8000 inbound to soma-inference-node tagged VMs
- Add soma-node-secret to Secret Manager (GCE startup script auth)
- Inject SOMA_NODE_SECRET, GCP_PROJECT_ID, GCP_ZONE, SOMA_PUBLIC_URL into
  the soma Cloud Run service via env vars
2026-04-28 10:48:32 -05:00
Will Anderson f2b025a433 Deploy soma to GCP Cloud Run at ai.neurontechnologies.ai
- Add cloud-run-soma.tf: soma-prod-us Cloud Run service in us-central1,
  neuron-soma-sa service account, soma Artifact Registry repo, Secret
  Manager secrets for HF token and operator key, serverless NEG, backend
  service, SSL cert
- Add dns-gcp.tf: Cloudflare A record for ai.neurontechnologies.ai pointing
  to GCP LB IP; Cloudflare provider added to main.tf/variables.tf
- Update load-balancer.tf: soma host rule + path matcher, soma SSL cert
  added to HTTPS proxy
- Update outputs.tf: soma service URL and artifact registry URL outputs
- Remove legion soma k8s manifests (Legion is gone)
- Update AGENTS.md to reflect GCP as primary production environment
2026-04-28 10:15:15 -05:00
Will Anderson 55a38bb693 fix: update Gitea URL to new GCP cluster instance
Legion hardware died April 27, 2026. New GCP k3s instance at 104.197.5.199.
Gitea now at gitea.git.svc.cluster.local:3000 within cluster.
2026-04-27 16:20:46 -05:00
Will Anderson fdbbfdb0ad iam: grant ci-pusher serviceAccountUser on all Cloud Run runtime SAs 2026-04-26 11:18:50 -05:00
Will Anderson e22dce6b5a iam: grant ci-pusher run.developer for Cloud Run deploys 2026-04-26 10:53:21 -05:00
Will Anderson 5b16aebdcb Remove stage environment from GCP — staging is local only 2026-04-26 01:09:44 -05:00
Will Anderson 53f2ad3c16 Fix Cloud Run probe ports and LB timeout_sec
- accounts: add ACCOUNTS_PORT=8080 env var (service defaults to 7753)
- api: add SERVER_PORT=8080 env var, change probe path to /actuator/health
- LB backends: remove timeout_sec (not supported for serverless NEG backends)
2026-04-26 00:28:33 -05:00
Will Anderson 36cfd3738d Fix Cloud Run: remove reserved PORT env, increase startup probe tolerance 2026-04-25 23:55:33 -05:00
Will Anderson bb583e3ccb Fix HCL syntax errors in accounts and api Cloud Run definitions 2026-04-25 22:54:18 -05:00
Will Anderson d4c65d5857 Expand GCP infra: accounts + API services, Cloud SQL, Artifact Registry
Architecture: intelligence stays on Legion; only compiled artifacts cross
to GCP. Source code and Neuron's knowledge base never leave the system.

Artifact Registry:
- neuron-marketing, neuron-accounts, neuron-api repos in us-central1
- Keep-last-10 cleanup policy; ci-pusher SA with writer access
- Legion CI runners authenticate via GCP_SA_KEY Gitea secret

Cloud SQL (cloud-sql.tf):
- postgres-15 on db-g1-small, us-central1 (scale up to REGIONAL HA at 1k users)
- Point-in-time recovery, 14-day backup retention
- Accounts DB + user; password generated and stored in Secret Manager
- JWT signing key in Secret Manager (shared by accounts + api)
- Cloud Run connects via built-in Auth Proxy (Unix socket volume mount)

Accounts Cloud Run (cloud-run-accounts.tf):
- 3 regions (us-central1, europe-west1, asia-northeast1), min:1 max:50
- Cloud SQL proxy volume mount; secrets via Secret Manager
- Stripe + JWT env vars; health probe on /health

API Cloud Run (cloud-run-api.tf):
- 3 regions, min:1 max:100, cpu_idle=false (always-hot)
- Validates JWTs from accounts service; no direct DB connection
- License admin token from Secret Manager

Load balancer (host-based routing):
- Same global anycast IP for all three services
- URL map routes by Host: neurontechnologies.ai→marketing,
  api.neurontechnologies.ai→api, accounts.neurontechnologies.ai→accounts
- New managed SSL certs for api.* and accounts.* added to HTTPS proxy
- Cloud Armor (WAF + rate limit) applied to all backends

Service accounts + IAM:
- neuron-accounts-sa: secretmanager.secretAccessor + cloudsql.client
- neuron-api-sa: secretmanager.secretAccessor
- allUsers invoker on all prod Cloud Run services (LB health checks)

bootstrap.sh:
- One-shot setup: pulls Stripe secrets from Vault → Secret Manager,
  creates CI SA JSON key, prints DNS + next-step instructions
2026-04-25 22:54:18 -05:00
Will Anderson 91bfc42989 ops: scale down all non-critical services — dedicate resources to Neuron
Scaled to replicas: 0: fornax (coordinator/ui/workers), jellyfin via media,
overseerr, plex, radarr, sonarr, bazarr, prowlarr, flaresolverr, ollama,
harmonic-wordpress, docuseal, listmonk, memos, plane, redpanda, wp-coordinator
2026-04-24 20:07:41 -05:00