search

mSvcBench / muBench

µBench is a tool for benchmarking cloud/edge computing platforms that run microservice applications. The tool creates dummy microservice applications that can be customized by the user and run on Kubernetes.
Other
42 stars 22 forks source link

difference between request processed in Runner.py with traces shows in Jeager Ui #18

Closed hassanSattariNia closed 5 months ago

hassanSattariNia commented 6 months ago

i try run Runner.py with 1000 event workload and run success like this picture: image but when i try see this request and result as json file include traces of my request

i just see 24 request trace , and this is large difference between them (1000 && 24): image

can any body tell me about difference of these numbers?

andreadetti commented 6 months ago

Jaeger does not track all requests, so you see 24 instead of 100. To see spam, check related boxes.

hassanSattariNia commented 6 months ago

in top problem i execute 1000 request and jeager just record 24 request. i checked monitoring config and i find out in this project , for monitoring used istio . istio have default rate sample 1% with this environment variable PILOT_TRACE_SAMPLING . i solve my problem with change monitoing-install.sh with bottom command helm install istiod istio/istiod -n istio-system --wait --set pilot.traceSampling=100 .

whats your idea about my work i do it ? is it correct way ?

andreadetti commented 5 months ago

It sounds good