Observability is really a developer's responsibility not that of the ops team.
3 replies
It's important to add a per-transaction ID to your logs at the beginning of a
request which gets reused in logs from underlying services to find logically
related events.
Important when looking at metrics: learn to filter out noise!
Introducing service meshes allows to instrument services for observability and
also control behaviour. AWS AppMesh provisions the open-source
@EnvoyProxy to wrap microservices
transparently.