AnyEvent::Monitor - Service Monitoring using AnyEvent
use AnyEvent::Monitor;
my $foo = AnyEvent::Monitor->new(
name => 'foo',
on_softfail => sub {
warn "==> service fail: $_[1]";
},
on_hardfail => sub {
my ($resume_check);
warn "==> service fail, should attempt to do something to fix it: $_[1]";
$resume_check->(60); # resume checking after 60 secs
},
on_resume => sub {
my ($prev, $outage) = @_;
if ($prev) {
warn "service resumed from: $prev, total outage: $outage secs";
}
});
$foo->install_timers( 300 ); # delay checking for 300 secs
sub my_polling_check {
my ($timestamp, $status) = @_;
# $foo->heartbeat($timestamp, $status);
}
$foo->status; # expecting "normal"
AnyEvent::Monitor provides a simple way to do periodical checks on given services, and provides callback when the service fails that you can attempt to fix it programmatically.
The callback to be called after service remains failed for $soft_timeout.
The callback to be called after service remains failed for $hard_timeout. You should attempt to fix the service and call $resume-($delay)> after the attempt has been made. This will make the monitoring resume after $delay seconds.
The callback to be called after service monitoring resumes. If it had failed, $previous_status and $outage seconds will be given.
Set the next checking timer according to soft_timeout and hard_timeout, with additional $delay from now. You don't normally need to call this method manually, unless you want to delay the start of the monitoring.
This is used to update the status of the service. only normal is meaningful to AnyEvent::Monitor. Other values are considered as the service failed.
Chia-liang Kao <clkao@clkao.org>
This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.
