Observability gaps when migrating from monolith to microservices

We're mid-migration from a monolith to microservices (Kubernetes, ~12 services so far). The biggest surprise has been how much observability we lost in the transition. In the monolith we had one set of logs, one tracing context, one error dashboard. Now we have fragmented logs across pods, broken trace propagation in 3 services, and no single pane for error rates. What observability stack did your teams standardize on during similar migrations? OpenTelemetry + Jaeger? Datadog? Something else? Bonus: how did you handle trace context propagation across async message queues (we're using RabbitMQ).