Fix for OpenReplay /v2/api/* 404 errors after running openreplay -u - patches ingress to route to correct backend service

openreplay-post-upgrade-fix.md

OpenReplay Post-Upgrade Fix: Health API 404

Problem

After running openreplay -u, the /v2/api/* endpoints return 404 because the ingress gets reset to default values that point to the wrong backend service.

Root Cause

• The api-openreplay ingress routes to api-openreplay:8080 (Go server)
• But the actual API (health, signup, etc.) lives on chalice-openreplay:8000 at /api/*
• The rewrite target also needs to be /api/$1 instead of /$1

Fix Commands

Run these after every openreplay -u:

# Point ingress to chalice service
kubectl patch ingress -n app api-openreplay --type='json' -p='[
  {"op": "replace", "path": "/spec/rules/0/http/paths/0/backend/service/name", "value": "chalice-openreplay"},
  {"op": "replace", "path": "/spec/rules/0/http/paths/0/backend/service/port/number", "value": 8000}
]'

# Fix rewrite target
kubectl annotate ingress -n app api-openreplay nginx.ingress.kubernetes.io/rewrite-target='/api/$1' --overwrite

Verify

# Test health endpoint
curl -s "https://replay.example.com/v2/api/"
# Expected: {}

# Test signup endpoint
curl -s "https://replay.example.com/v2/api/signup"
# Expected: {"data":{"tenants":true,...}}

Notes

• The health endpoint is at /api/ (root), not /api/health
• Chalice is a Python/FastAPI service serving the main API
• The Go api-openreplay service serves different functionality