$ date -u; podman stop -t 10 messenger-lab3_api_1; date -u
02:07:18.261
02:07:21.454

$ podman logs messenger-lab3_api_1 (shutdown sequence, UTC)
02:07:18.345 shutdown signal received, draining 3s 5s
02:07:21.345 server stopped  

$ TZ=UTC podman logs -t messenger-lab3_lb_1 (backend state, UTC)
02:07:20.346 Server api/api2 is DOWN, reason: Layer7 wrong status, code: 503, info: "Service Unavailable", check duration: 0ms. 1 active and 0 backup 

SIGTERM 02:07:18.345 -> LB marks DOWN by 503 readiness 02:07:20.346 (2.0 s, <= inter 1s x fall 2) -> listener closed 02:07:21.345 (drain 3 s); hook returned in 3.2 s.
HAProxy slot names (api1..api4 from server-template) are assigned by DNS order and do not match container numbers.
