We’ve recently been seeing poor P99 edge performance on our main app, despite app-level response times being consistently low.
Has this issue been identified/is there a resolution timeline?
There was increased packet losses between our MIA edges and MIA worker servers. This caused a lot of TCP retransmission and slower response times. Should have been fixed a few hours ago when our network provider re-routed traffic.
This also happened yesterday morning and a similar change had to be implemented.
There should’ve been a status page update. I’ll add 2 updates retroactively.
Thanks for the info. Is there something we can do on our end to reduce the likelihood of these issues in the future? It sounds like this was specifically related to MIA edge to MIA server traffic, so I assume that not having app servers deployed in this region would have helped?
I wonder if this is something Fly’s edge could detect and route around to a backup region.