Valentín Gutiérrez

Things have improved a lot by backporting https://github.com/apache/trafficserver/pull/5697, I'm not sure it's the only one, but that's definitely helping:)

The issue is still there sadly : 

perf top points to an issue related to HTTP/2: ``` 10.12% traffic_server [.] Http2Stream::signal_write_event 10.01% traffic_server [.] Http2Stream::update_write_request 9.98% traffic_server [.] Http2ConnectionState::send_a_data_frame ```

A service discovery client in form of a CLI tool could be pretty useful for debugging purposes, `ping $(discovery service1-backend)` or `nc -zv $(discovery service1-backend) 443`. It can be easily...

@svinota any progress on `fibmatch` support?

> Could you test the 2.6.16 ? It contains many fixes. It is hard to say if one of them concerns your issue. done: ```/usr/sbin/haproxy -vv HAProxy version 2.6.16-1~bpo11+1 2023/12/13...

@capflam issue is still there with 2.6.16: ``` Running tasks: 20 (48 threads) function places % lat_tot lat_avg ssl_sock_io_cb 20 100.0 - - ```

@capflam I got some task activity details via `show profiling tasks`: ``` Tasks activity: function calls cpu_tot cpu_avg lat_tot lat_avg ssl_sock_io_cb 327775089 5.180m 948.0ns 16.70m 3.057us h2_io_cb 772000 10.65s 13.80us...

we are still seeing this in production, @capflam any tips on what to check that could help moving this issue forward?

comparing `strace -k` output between a healthy server and the impacted one there is one stack trace that doesn't show on the healthy one (0 occurrences on the healthy VS...