[google-appengine] Re: Debugging large gap between appLatencySeconds and latencySeconds

2020-09-17 Thread 'Juan Carlos Gomez' via Google App Engine
Hello, appLatencySeconds and latencySeconds are equal to nginx variables. With, appLatencySeconds == $upstream_response_time [1] and latencySeconds == $request_time[2]. For clarity I'll include the definitions: appLatencySeconds :$upstream_response_time: keeps time spent on receiving the

[google-appengine] Re: Debugging large gap between appLatencySeconds and latencySeconds

2020-09-16 Thread Eddie Zhou
Sorry -- correction, our target cpu utilization is at the default of 0.5, but the utilization at the time was very much below that. On Wednesday, September 16, 2020 at 11:55:37 AM UTC-7 Eddie Zhou wrote: > We were actually able to find the Load Balancer logs (hoping this helps > someone else

[google-appengine] Re: Debugging large gap between appLatencySeconds and latencySeconds

2020-09-16 Thread Eddie Zhou
We were actually able to find the Load Balancer logs (hoping this helps someone else in the future) via resource.type="http_load_balancer" and connecting trace to jsonPayload.trace from App Engine nginx.request. In our case, this shows that the load balancer received the client request at