The webhook handler's client had no timeout, and slog calls the handler
inside the log call, so a server that accepted the connection and never
answered stopped that log call for good, and every later one. A request
still running after 5 seconds, reading the answer included, now fails
with a timeout error. The handler also reads the answer to the end
before closing it, so the connection is reused for the next record.
Two new tests point the handler at a server that never answers and at
one that sends its status and then stalls the answer, and check that
Handle returns a timeout error within the timeout. The README states
the timeout.
Model: opus-5-5