Request tracing via Render's Load Balancer
F
Fergal Armstrong
Seeing tracing data from the load balancer would give a more complete picture. I can see the time spent in services, but sometimes our service doesn't work, or it takes a long time for the load balancer to reach the service and our latency recorded isn't accurate.
I am imagining this as being able to configure an opentelemetry collector endpoint for the account or per service, and the load balancer sends spans, starting a new trace and adding headers if needed (maybe one day the CDN can forward tracing too and traces will start there).
Log In
R
Rich Shakespeare
We think this is vital to be able to see 5xx errors and to know how to adjust scaling/app perf. There's no visibility at the moment at all if anything is amiss