High Average Response Time

Question for the community. I included a screenshot of my API chart. I am wondering why my average response would be so high or how I might go about tracking that down? The services that are hosted on this guy dont seem slow. However is there perhaps a request in there somewhere that is running slow and often that might be causing this? Is that tracking Client->Traefik->Server->Trafik->Client for a round trip or how is that field calculated?

You can hook into OpenTracing (check out Jaeger or Zipkin) to get greater insight on where the delay is

1 Like

Take a look at your access log. I noticed the same thing on my end, and found in access.log that most requests were taking 1-5ms, but then found that a single API call from the app I’m testing that happened to take a very, very long time (10000+ms) and it was driving the average up much higher.

It might not be Traefik that’s driving the average response time up, but the server itself. At least it was in my case, and I haven’t yet found a way to hide specific request paths from the dashboard’s average tracking.

I fired up Jaeger and was able to find that there is a part of out management software that is holding open a websocket type connection for 60-120s at a time. Thanks @jak