Hi,
My app (Rails web app, deployed SJC) suddenly started seeing spikes in Request Queueing times in SJC area starting this week. Here’s a breakdown of request queueing averages:
I thought it might be something about anycast routing (full disclosure: I don’t understand much) from different entry nodes, so I looked at request country codes to see if there was any clear separation between different regions. Request queuing appears pretty bad across the board:
The strange thing is that personally / anecdotally I do not experience any slowdowns (I am in LAX region, and sending requests to my server in SJC). So I am unable to reproduce on my end.
The final thought is that perhaps this is simply a change in measurement methodology and the way that Fly is reporting metrics has changed recently.
Anybody experience the same on their end?