16 of 20 rust services were crashing on boot because their main.rs calls
std::env::var('PORT').expect(...). This commit adds the missing PORT env
to each deployment YAML, matching its containerPort (9100-9118).
Also adds BEECEPTOR_URL to the ConfigMap - the payments service requires
it for the mock payment gateway integration.
Adds apps/registry/ with retention script + CronJob (keep last 3 SHA
builds, preserve -latest aliases) to prevent the single-build wipeout
that caused the original registry outage.
AI assistant image also rebuilt: 2876b45 (main branch) - it was on a
ghost SHA tag that was GC'd.
16 of 20 rust services had no PORT env var set; their main.rs calls
std::env::var('PORT').expect('PORT must be a valid u16') which panicked
on startup. This commit adds env.PORT matching the existing containerPort
for each service. Service ports: gateway=9100 users=9101 companies=9102
jobs=9103 job_seekers=9104 customers=9105 employees=9106 photographers=9107
tutors=9108 makeup_artists=9109 developers=9110 video_editors=9111
graphic_designers=9112 social_media_managers=9113 fitness_trainers=9114
catering_services=9115 payments=9116 ugc_content_creators=9117 leads=9118
- Backend: 9444056 - crane mirror base images, then build
- Frontend: d26f0bf - crane mirror node:20-alpine, then build
- Admin: e044d4c - crane mirror node:20-alpine, then build
- AI: 6f94539 - crane mirror base images, then build
- All repos now use crane to mirror base images before building
- frontend: use registry.nxtgauge.com/node:20-alpine
- admin: use registry.nxtgauge.com/node:20-alpine
- ai: use registry.nxtgauge.com/rust:1.87-alpine and alpine:3.20
- All images now use internal registry to avoid Docker Hub rate limits
- Update kustomization.yaml tags and admin base deployment image
- Remove ghcr.io references from admin gitops
Images needed in registry.nxtgauge.com:
- node:20-alpine
- rust:1.87-alpine
- alpine:3.20
- Change ghcr.io/traceworks2023/nxtgauge-db-migrate to registry.nxtgauge.com/nxtgauge-db-migrate
- Ensure all images in gitops reference registry.nxtgauge.com
- Change frontend image from ghcr.io to registry.nxtgauge.com
- Remove duplicate frontend-solid entry in backend-rust kustomization
- Update frontend tag to 80fb6aa with multi-stage Dockerfile fix
- Add namespace, PVC, deployment, service, and kustomization for Ollama
- ClusterIP only (internal), no ingress
- 10Gi PVC mounted at /root/.ollama
- Resource limits: 700Mi-1500Mi memory, 500m-1000m CPU
- Add ArgoCD Application CR for ollama
- Update backend configmap with OLLAMA_BASE_URL and OLLAMA_CHAT_MODEL
- Add nxtgauge-ai-assistant deployment, service, and kustomization overlays
- Update users service to commit 430711a with AI endpoints
- Set NXTGAUGE_USERS_URL and AI_SERVICE_KEY env vars
