Hi,
I need advice to understand the average response time given by Traefik metrics (Prometheus)
I have created a simple FASTApi project, with an /health a /ping and hello world and response time given by Traefik is around 11 milliseconds.
With a browser or curl it is more around 45 ms.
curl -w "@curl-format.txt" -o /dev/null -s "http://fastapi-sample.service.gra.dev.consul/v1/ping"
time_namelookup: 0.018462s
time_connect: 0.027073s
time_appconnect: 0.000000s
time_pretransfer: 0.027103s
time_redirect: 0.000000s
time_starttransfer: 0.047565s
----------
time_total: 0.047613s
I can see in Traefik logs "duration": 0.010641353000000001 (in nanoseconds) so 10.64 ms
Traefik prometheus configuration is the default
[metrics.prometheus]
entryPoint = "metrics"
addRoutersLabels = true
# buckets = [0.1,0.3,1.2,5.0]
As the lower bucket is 0.1 I assume that leverage will never go lower than 10ms
I have also seen other requests like Apache /server-status be at 4ms
Is it fair to have such a low duration given by Traefik?
Why Traefik do not give me the same response time as my browser or Curl?
Thank you in advance for enlightening my ignorance.