Juno - Asynchronous event-driven checking mechanism
version 0.010
This runs an asynchronous checker on two servers (jack and jill), running an HTTP test every 10 seconds with an additional Host header.
my $juno = Juno->new(
hosts => [ 'jack', 'jill' ],
interval => 10,
checks => {
HTTP => {
headers => {
{ 'Host', 'example.com' },
},
on_result => sub {
my $result = shift;
...
},
},
},
);
# makes juno run in the background
$juno->run;
Juno is a hub of checking methods (HTTP, Ping, SNMP, etc.) meant to provide developers with an asynchronous event-based checking agent that has callable events for the each check's result. You can then use this probed data for whatever it is you want. This helps you write stuff like monitoring services, amongst other things.
Juno is flexible, introspective and composable but still very straight-forward. To use Juno, you simply need to create a new object, give it all the information needed (such as the test you want to run) and the optional (and possibly required) additional check-specific arguments, and then just let it run in the background, working for you.
Each check Juno runs can have multiple events launched for it. The events will only be called if you ask them to be called. This means that if you do not want a certain event to run for a specific check, just don't provide it.
Let's go over the events you can run. Each check should provide all of these events unless specified otherwise in the check's documentation.
Also, note that each check tried to decide if a result has been successful or not. This could be decided using return codes, response headers, failed results, etc. You have callbacks for successful results or not, but they do not prevent you from using the other callbacks provided. They will still work seamlessly if you provide all of them.
on_before
This event gets called before the check is actually called.
One thing you can do with it is to check how long a request takes. You can timestamp before a check is done, and once a check has a result, you can timestamp that again. The diff between those timestamps is how long it took to run the check.
on_result
This event gets called as soon as a check result has come in. It might have failed, it might have been successful. It doesn't matter.
One usage for it is to timestamp the result to correspond to the
before
callback explained above.Another usage for it is in case you want to check for yourself if the result has been successful or not. Perhaps you have your own API you're testing against and you (and only you) can decide if it was successful.
You will receive the entire result and can make any decision you want.
on_success
This callback is called when the check decides the result has succeeded.
on_fail
This callback is called when the check decides the result has failed.
Hopefully by now you understand the basic concepts of Juno. The following documentation will help you get started on how to use Juno exactly, creating a new object and providing it with all the information required.
The checks you want to run.
This is a hashref of the checks. The key is the check itself (correlates to the class in Juno::Check::
) and the values are the attributes to that check.
my $juno = Juno->new(
hosts => [ '10.0.0.2', '10.0.0.3' ],
checks => {
HTTP => {
path => '/test',
},
},
);
The checks
argument is the most important one, and it is mandatory. This defines what will be checked, and adds additional parameters. Some might be optional, some might be mandatory. You should read each check's documentation to know what options are available and which are required.
If you need to run multiple checks of the same type, such as two different HTTP tests, you will need to run two Juno instances. It's perfectly fine, because Juno has no global variables and works seamlessly with multiple instances.
Hopefully this will change in the future, providing more advanced options to have multiple checks of the same type.
An arrayref of hosts you want all checks to monitor.
my $juno = Juno->new(
hosts => [ '10.0.1.100', 'sub.domain.com' ],
...
);
The interval for every check.
my $juno = Juno->new(
interval => 5.6,
...
);
This sets every check to be run every 5.6 seconds.
Default: 10 seconds.
Delay seconds for first check.
my $juno = Juno->new(
after => 10,
...
);
This will force all checks to only begin after 10 seconds. It will basically rest for 10 seconds and then start the checks. We can't really think of many reasons why you would need this (perhaps waiting for a database connection?), but nonetheless it is an optional feature and you should have control over it if you want to change it.
If this is set to zero (the default), it will not delay the execution of the checks.
Default: 0 seconds
The prop_attributes
are an arrayref of attributes that are propagated from the main Juno object to each check object. This could be a hard-coded list, but it's cleaner to put it in an attribute. This means it's available for you to change. There really is no need for you to do that.
my $juno = Juno->new(
prop_attributes => [ 'hosts', 'interval' ],
...
);
Default: hosts, interval, after.
Run Juno.
use Juno;
use AnyEvent;
my $cv = AnyEvent->condvar;
my $juno = Juno->new(...);
$juno->run;
$cv->recv;
When you call Juno's run
method, it will begin running the checks. Separating the running to a method allows you to set up a Juno object (or several Juno objects) in advance and calling them later on when you're ready for them to start working.
However, note that running Juno will not keep the program running by itself. You will need some condition to keep the program running, as demonstrated above.
Sawyer X <xsawyerx@cpan.org>
Adam Balali <adamba@cpan.org>
This software is copyright (c) 2012 by Sawyer X.
This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.