bang: detect SIGINT and SIGTERM failures
Detect bang failures because the process was killed with SIGINT or
SIGTERM and count those as a failure.
This way updates that start failing because they are now hanging every
time (and then being killed by jobd.pl) will be caught and generate a
failure message after the minimum count/interval is reached rather than
silently continuing to fail with no notification.
Previously this detection was undesirable as it would have likely
resulted in additional spurious failed+recovery messages. However, now
that there is a minimum failure count and/or interval required including
SIGINT and SIGTERM kills as failures will no longer cause that problem.
Signed-off-by: Kyle J. McKay <mackyle@gmail.com>