5.5 — Configure auto-scaling rules

[davidortinau]

Description

Configure Container Apps auto-scaling rules based on resource usage and HTTP metrics. Automatically add/remove replicas to handle load spikes and reduce costs during quiet times.

Dependencies

• Depends on 3.5 — Configure Azure Container Apps for services #52 (Container Apps)
• Depends on 5.1 — Enable Azure Monitor and Application Insights #60 (Application Insights)

Acceptance Criteria

• Scaling rules configured for each service:
  • API: Scale on HTTP concurrent requests (>50 per instance)
  • WebApp: Scale on CPU (>70%)
  • Workers: Scale on queue depth (if applicable)
  • Marketing: Fixed 1 replica (no scaling)
• Min/max replica settings:
  • Staging: 1-2 replicas
  • Production: 2-5 replicas (minimum 2 for HA)
• Tested: scale-up triggered under load, scale-down when idle
• Cost optimized: appropriate SKU and scaling ranges
• Dashboards show scaling events

Technical Notes

• HTTP-based scaling is most common for web services
• CPU-based scaling for compute-bound services
• Queue-based scaling for background workers
• Scale-down requires grace period to drain connections

---

Phase: 5 | Size: M | Owner: Zoe (Lead)