Microservices Challenges — boundaries, failures, observability

Boundaries

• Model capabilities, not tables
• Minimize shared databases; use contracts and events
• Keep cross‑service communication explicit

Failures

• Circuit breakers and timeouts
• Bulkheads per adapter to isolate issues
• Dead‑letter queues for poison payloads

Observability

• Structured logs with correlation IDs
• Metrics for latency, error rate, saturation
• Traces across critical flows