fetch2 timeout too long / not working

[ldemailly]

When the host/ip isn't reachable it seems it hangs indefinitely or at least for a very long time

[ldemailly]

On mac, added to /etc/pf.conf

block drop out proto tcp from any to 18.222.136.83 port 443

and

sudo pfctl -f /etc/pf.conf

So the first address for https://debug.fortio.org/ is blocked

Updated #616 as noticed that the std client which is now the default in the UI gets stuck on first IP even with 3 or 6 connections - but does have a timeout, so does fortio curl is also timing out as expected (unless using -dns-method rnd then it timesout 1/3rd of the time when using that blocked ip/port)

And indeed fetch2 handler blocks until you navigate away and only then one gets

13:20:28.592 r41 [ERR] http_server.go:541> Error for "https://debug.fortio.org/": Get "https://debug.fortio.org/": context canceled

[ldemailly]

went back to go 1.19 and fortio 1.40.1 and same behavior, so not a regression but never handled hanging connection in proxy