High API latency over both HTTP & HTTPS

[jochenbillen]

HTTP roundtrip times have always have been in the 400ms region. However, since Friday May 16th, we see persistent RTT up to 7000ms causing a massive amount of queued events on our back-end systems.

We are Europe based and we measure HTTP RTT from both our private datacenters as well as from the AWS datacenter in Frankfurt Germany.

Is this a known issue? Or rather a regional problem?

[jkraybill]

We have been seeing this issue on/off for as long as you have. I've sent three emails to Stathat support with no response, so for now we are working around this by swallowing timeouts, but we will have to port to a different stats service if this is not resolved.