Rolling out distributed tracing for the Cloud Migration Platform

Integrated OpenTelemetry across all services in the Cloud Migration Platform. We're now collecting traces end-to-end from the API gateway down to the database layer.

Context

The Cloud Migration Platform had fragmented monitoring — separate tools for logs, metrics, and APM. Debugging cross-service issues required correlating data from three different UIs.

What We Did

We rolled out OpenTelemetry instrumentation across all 12 microservices:

Auto-instrumentation for HTTP, gRPC, and database calls
Custom spans for critical business logic paths
Context propagation through SQS/SNS message headers

Results

After one week of production data:

Average time to diagnose issues dropped from 45 min to 8 min
Identified 3 previously unknown N+1 query patterns
Trace sampling at 10% keeps storage costs under control

Next Steps

We’re evaluating Grafana Tempo as a drop-in replacement for our current tracing backend to reduce costs further.