After migrating 14 production microservices across 200 hosts from Datadog 7.0 to a stack built on OpenTelemetry 1.20 and Prometheus 2.50, my team cut observability spend by 72% ($21k/month to $5.8k/month) with zero loss in metric fidelity, and 18% faster query performance for p99 traces. If you’re still paying Datadog’s per-host, per-custom-metric...