Ive set up a basic opentelemetry pipeline using Docker, Jaeger, Prometheus and Grafana to examine the data. For reasons I dont yet understand my Jeager traces are being sliced into blocks with only the last block presented.
The bahviour indicates that traces are being buffered and removed from Grafana charts according to either Jaeger or Grafana configuration. I suspect (though not sure) that this is Jaeger configuration so Im seeking a way to increase the trace buffer size either in time or magnitude.
You can see the effect here. When I run this the traces are being sliced into blocks and removed as each new block is presented.
There is a working github repo here. Clone, install (npm i) and docker compose up -d to try it out. Further instructions on accessing and starting traces in the readme.
I tried setting the JAEGER_REPORTER_MAX_QUEUE_SIZE in my docker compose file but this had no effect. i.e.
jaeger:
image: jaegertracing/all-in-one:latest
container_name: jaeger
ports:
- '16686:16686'
- '14268'
- '14250'
environment:
- JAEGER_REPORTER_MAX_QUEUE_SIZE=10000
My question is: how can i increase the length of my Jaeger traces so that I can view a full import and cross correlate to ALL my metrics.
