Cody Jarrett

It looks like the "manual" routes (which we don't really use) loads first and so the listeners go active before the full list of services comes back from Consul services.

Having this same issue on 1.5.11 and 1.5.13. Just random 502s in load test, like .8-1.5% of traffic. If I hit the app container directly, no errors. If I test...

I took a packet capture between fabio and the backend container. When the 502 happens, fabio sends a SYN to the backend, the backend responds with a SYN ACK, and...

I wanted to circle back here and say that our intermittent 502s were actually app problems with the backends that fabio was routing to. In more packet captures, we realized...

I have an example use case here as well. We have several thousand services and about 7k routes. When a fabio restart (or starts up in the case of a...

Hi @leprechau, What samples can I provide here? More CPU charts helpful? The left side of that cpu graph is when running on 1.5.11, and the right side is when...

I should also note that when I queried services against the consul agent on the host where the problematic fabio was, I was getting an accurate list of services and...

We ran a script across the environment to pull the services from consul, then query every fabio and compare the routes in fabio to consul services destinations. We still have...

I hit the `:9998/health` endpoint and it returns `OK`. Is there anything else I can look at? Here's from the docker logs on one host. Around 02:49:54, the log pattern...

Ok very interesting. I'll look into your observations. I hit that consul health/state/any endpoint on this host where the fabio container with stale routes is running. I see my specific...