chore: release v1.0.7

Playout: slate fallback, strict ffprobe validation, Prometheus metrics
Chat: history on join, DM offline delivery, env-config rate limits,
  Redis auto-reconnect, heartbeat, ban/mute enforcement, DM metrics
Viewer: presence indicators, schedule overlay, public /api/schedule
Admin: playlist drag-reorder, schedule overrides CRUD, user actions,
  inline mod buttons, CSV export, config editor, restart buttons,
  upload normalization gate, analytics time range
nginx: CORS preflight, JSON health, stub_status
Docs: test-lab vs production distinction across all configs/docs
Infra: Redis maxmemory, CD health IP-agnostic, slate volume mount

.gitea/workflows/cd.yml

@@ -1,3 +1,8 @@
+# CD workflow — TEST LAB DEPLOYMENT
+# Deploys to test lab host 192.168.88.40:2375.
+# Production Hetzner target TBD — switch DOCKER_HOST when provisioned.
+# See TODO.md Production Deployment section for cutover plan.
+
 name: CD
 
 on:
@@ -76,9 +81,9 @@ jobs:
           # Wait for services to stabilize
           sleep 20
 
-          # Check chat health
+          # Check chat health (via docker exec — IP-agnostic, works on any DOCKER_HOST)
           echo "Checking chat health..."
-          if ! curl -f http://192.168.88.40:3000/health; then
+          if ! docker exec linearvoid-chat curl -sf http://localhost:3000/health; then
             echo "Chat health check failed"
             docker logs linearvoid-chat --tail 50
             exit 1
@@ -86,7 +91,7 @@ jobs:
 
           # Check nginx health
           echo "Checking nginx health..."
-          if ! curl -f http://192.168.88.40/health/; then
+          if ! docker exec linearvoid-nginx curl -sf http://localhost/health; then
             echo "Nginx health check failed"
             docker logs linearvoid-nginx --tail 50
             exit 1