1

I am making use of web socket functionality on GAE flexible environment to enable a GraphQL Subscription on our API. However I have noticed in my google cloud reporting there are massive latency spikes of 30+ minutes. See below: enter image description here

Here is the log related to this requuest enter image description here

I can see it is status 101 which is usually related to a websocket. Looking in the chrome debugger when using the application that accesses the API I can see there is a pending web socket which indeed has a large request time as essentially the duration of the website visit. enter image description here

Is this the expected behaviour of the latency reporting? It doesn't seem to affect performance, however it makes it difficult to see what the real latency of our normal API requests are.

Thanks!

Alex Fox
  • 1,175
  • 2
  • 15
  • 27

1 Answers1

0

I found an iteration of this 101 requests being related to changes on the protocols and ports being performed and thus, as the connections are not being closed and stay in pending, this originates the latencies you experienced. [1] [2]

This makes me think the root cause of this issue may be due to your configuration of GraphQL. [3]

Roger
  • 158
  • 8