view release on metacpan or search on metacpan
Makefile.PL
README
README.md
bin/standby-mgm-cgi.pl
bin/standby-mgm.pl
bin/standby-mgm.psgi
conf/standby-mgm.conf
conf/standby-mgm.conf.dist
dist.ini
examples/complex/README
examples/complex/complex_endpoint.pl
examples/complex/lib/App/Standby/Cmd/Command/complex.pm
examples/complex/lib/App/Standby/Service/Complex.pm
examples/simple/README
examples/simple/lib/App/Standby/Service/Simple.pm
examples/simple/simple_endpoint.pl
lib/App/Standby.pm
lib/App/Standby/Cmd.pm
lib/App/Standby/Cmd/Command.pm
lib/App/Standby/Cmd/Command/bootstrap.pm
lib/App/Standby/DB.pm
lib/App/Standby/Frontend.pm
lib/App/Standby/Group.pm
lib/App/Standby/Service.pm
lib/App/Standby/Service/HTTP.pm
lib/App/Standby/Service/MS.pm
NAME
App::Standby - Managing on-call rotation and notification queues
DESCRIPTION
This distribution provides a small Plack webapp which helps with
managing on-call rotations and notification queues. It allows you to
manage several different queues from on place. It is easily extendible
by plugins which can talk to virtually API endpoint to update a queue or
a contact.
Most organizations have at least one big monitoring system (like Nagios
or Zabbix) and at least one external service level monitoring and other
means of notification, If you don't want to pass around a shared on-call
mobile you have to remember to update all those services when the one on
duty changes. This app will help you with that.
It allows you to manage several groups with their own queues and update
each groups external services with just one click.
In order for this app to do anything there must be two things in the
database: users and services. This section will show you how to create
the later.
First select services and open up the new service dialog. Enter a short
name for this service. Remember it must be all lowercase alphanumerics
since it's going to be used as a prefix for the configuration values
later. The description can be anything. Select the appropriate class and
enter the group password you've dedfined when bootstrapping the service.
Choose class HTTP for a simple endpoint which just gets the whole
ordered user list as a JSON string. Chosse MS for a Monitoring::Spooler
endpoint and Pingdom if you have an Pingdom account.
Add as many services as necessary.
Next select config from the menu and add new config items. For each
service at least one. The HTTP service need an endpoint, so if the HTTP
service was called simple then the config item for the endpoint must be
called simple_endpoint and contain something like
http://simple.domain/api/.
For any pingdom service there must be at least four config items. The
necessary items are apikey, username, password and contact_id. The last
one may be given multiple times to update multiple Pingdom contacts in
one account. If your service is called pingdom the keys would be called
pingdom_apikey, pingdom_username, pingdom_password and
pingdom_contact_id.
The MS service class needs an endpoint and a group id (name_group_id)
which is used to update the appropriate group in the MS DB. If you call
your MS service "ms" then the necessary keys would be called
"ms_endpoint" with a value of e.g.
"http://localhost/ms/?rm=update_queue" and "ms_group_id" with a value of
e.g. "1".
Have a look at the CPAN distribution Monitoring::Spooler for more
documentation on MS.
CONTACTS
To be able to notify someone you must add some contacts. Since the whole
point of this app is to help with managing changing on call rotations
you should create more than one contact.
examples/complex/README view on Meta::CPAN
This example includes a more complex service example.
The service is just a subclass of App::Standby::Service::HTTP and
sends the whole queue. It also provides a cronjob which calls the
service once a day to keep your endpoint up to date.
The endpoint examines the request and updates your complex service.
examples/complex/lib/App/Standby/Cmd/Command/complex.pm view on Meta::CPAN
__END__
=head1 NAME
App::Standby::Cmd::Command::complex - Example for a command to be run as a cronjob
=head1 DESCRIPTION
This class implements an example for a cronjob that is run once a day to keep some
complex endpoint up-to-date.
=cut
examples/complex/lib/App/Standby/Service/Complex.pm view on Meta::CPAN
# use MooseX::Params::Validate;
# use Carp;
# use English qw( -no_match_vars );
# use Try::Tiny;
# extends ...
extends 'App::Standby::Service::HTTP';
# has ...
# with ...
# initializers ...
sub _init_endpoints {
my $self = shift;
return $self->_config_values($self->name().'_endpoint');
}
# your code here ...
no Moose;
__PACKAGE__->meta->make_immutable;
1;
examples/simple/README view on Meta::CPAN
This example includes a very simple service example.
The service is just a subclass of App::Standby::Service::HTTP and
sends the whole queue while the endpoint is implemented as
a very simple CGI script that decodes the queue and inserts it
into a queue in a MySQL database.
examples/simple/lib/App/Standby/Service/Simple.pm view on Meta::CPAN
# use MooseX::Params::Validate;
# use Carp;
# use English qw( -no_match_vars );
# use Try::Tiny;
# extends ...
extends 'App::Standby::Service::HTTP';
# has ...
# with ...
# initializers ...
sub _init_endpoints {
my $self = shift;
return $self->_config_values($self->name().'_endpoint');
}
# your code here ...
no Moose;
__PACKAGE__->meta->make_immutable;
1;
__END__
lib/App/Standby.pm view on Meta::CPAN
=head1 NAME
App::Standby - Managing on-call rotation and notification queues
=head1 DESCRIPTION
This distribution provides a small Plack webapp which helps with managing on-call rotations
and notification queues. It allows you to manage several different queues from
on place. It is easily extendible by plugins which can talk to virtually API
endpoint to update a queue or a contact.
Most organizations have at least one big monitoring system (like Nagios or Zabbix) and
at least one external service level monitoring and other means of notification,
If you don't want to pass around a shared on-call mobile you have to remember to update all
those services when the one on duty changes. This app will help you with that.
It allows you to manage several groups with their own queues and update each
groups external services with just one click.
=head1 METHODS
lib/App/Standby.pm view on Meta::CPAN
=head2 SERVICES
In order for this app to do anything there must be two things in the database: users and services.
This section will show you how to create the later.
First select services and open up the new service dialog. Enter a short name for this service.
Remember it must be all lowercase alphanumerics since it's going to be used as a prefix for the
configuration values later. The description can be anything. Select the appropriate class and
enter the group password you've dedfined when bootstrapping the service.
Choose class HTTP for a simple endpoint which just gets the whole ordered user list as a JSON
string. Chosse MS for a Monitoring::Spooler endpoint and Pingdom if you have an Pingdom account.
Add as many services as necessary.
Next select config from the menu and add new config items. For each service at least one. The
HTTP service need an endpoint, so if the HTTP service was called simple then the config item
for the endpoint must be called simple_endpoint and contain something like http://simple.domain/api/.
For any pingdom service there must be at least four config items. The necessary items are apikey,
username, password and contact_id. The last one may be given multiple times to update
multiple Pingdom contacts in one account. If your service is called pingdom the keys would be
called pingdom_apikey, pingdom_username, pingdom_password and pingdom_contact_id.
The MS service class needs an endpoint and a group id (name_group_id) which is used
to update the appropriate group in the MS DB. If you call your MS service "ms" then the necessary
keys would be called "ms_endpoint" with a value of e.g. "http://localhost/ms/?rm=update_queue" and
"ms_group_id" with a value of e.g. "1".
Have a look at the CPAN distribution Monitoring::Spooler for more documentation on MS.
=head2 CONTACTS
To be able to notify someone you must add some contacts. Since the whole point of this app
is to help with managing changing on call rotations you should create more than one contact.
Create at least two contacts. The name may be anything but you should know who it is refering to.
lib/App/Standby/Service.pm view on Meta::CPAN
=head2 update
This method is called with a array_ref containing the new ordering on any changed.
=head1 NAME
App::Standby::Service - Service Plugin baseclass
=head1 ADDING A NEW SERVICE
First of all there are two kinds of services: Simple HTTP endpoints and complex plugins.
The simple HTTP plugins just receive the whole queue in as JSON encoded array.
Those only need to subclass App::Standby::Service::HTTP and provide an implementation
for _init_endpoints. Have a look at the simple example.
All other services will need to subclass App::Standby::Service and implement an
update() method. Have a look at App::Standby::Service::Pingom for an example.
The method _config_values helps with getting values to known keys from
the config table. A service plugin MUST always prepend its name to
the key to allow for multiple instances of one plugin registered at
the same time.
=head1 AUTHOR
lib/App/Standby/Service/HTTP.pm view on Meta::CPAN
'isa' => 'Str',
'required' => 0,
);
has 'password' => (
'is' => 'rw',
'isa' => 'Str',
'required' => 0,
);
has 'endpoints' => (
'is' => 'rw',
'isa' => 'ArrayRef',
'lazy' => 1,
'builder' => '_init_endpoints',
);
# with ...
# initializers ...
sub _init_json {
my $self = shift;
my $JSON = JSON::->new->utf8();
return $JSON;
}
lib/App/Standby/Service/HTTP.pm view on Meta::CPAN
return $content;
}
sub _update {
my $self = shift;
my $user_ref = shift;
my $count = 0;
foreach my $endpoint (@{$self->endpoints()}) {
$self->logger()->log( message => "Updating endpoint: ".$endpoint, level => 'debug', );
my $req = HTTP::Request::->new( POST => $endpoint );
$req->content_type('application/x-www-form-urlencoded');
my $payload = $self->_build_payload($user_ref);
$req->content($payload);
$self->logger()->log( message => "Payload: ".$payload, level => 'debug', );
if($self->username() && $self->password()) {
$req->authorization_basic( $self->username(), $self->password() );
}
my $content;
my $response;
my $prev_alarm = 0;
my $success = try {
local $SIG{ALRM} = sub { die "alarm-standby-service\n"; };
$prev_alarm = alarm 10;
$response = $self->_ua()->request($req);
if ( !$response->is_success ) {
my $msg = "ERROR Request to $endpoint failed: " . $response->code() . ' - ' . $response->message();
$self->logger()->log( message => $msg, level => 'error', );
die( $msg );
}
$content = $response->content;
if ( !$content ) {
my $msg = "ERROR No content at $endpoint : " . $response->code() . ' - ' . $response->message();
$self->logger()->log( message => $msg, level => 'error', );
die( $msg );
}
1; # make sure $success has a true value ...
}
catch {
$self->logger()->log( message => "Request failed: ".$_, level => 'debug', );
}
finally {
lib/App/Standby/Service/MS.pm view on Meta::CPAN
# use MooseX::Params::Validate;
# use Carp;
# use English qw( -no_match_vars );
# use Try::Tiny;
# extends ...
extends 'App::Standby::Service::HTTP';
# has ...
# with ...
# initializers ...
sub _init_endpoints {
my $self = shift;
return $self->_config_values($self->name().'_endpoint');
}
# your code here ...
no Moose;
__PACKAGE__->meta->make_immutable;
1;
__END__
t/frontend.t view on Meta::CPAN
unlike($res->header('Location'), qr/Invalid...Key/, 'No error due to invalid key');
# contacts page for group 1 lists the new contact
$res = $cb->(GET '/?rm=list_contacts&group_id=1',);
like($res->content, qr/Testcontact/, 'User Testcontact exists in group 1');
# adding config item succeeds w/ correct key
$res = $cb->(POST '/', [
rm => 'insert_config',
'group_id' => 1,
'key' => 'ms_endpoint',
'value' => 'http://localhost/ms/',
'group_key' => 'test',
]);
ok($res->is_redirect, 'Redirect after create');
unlike($res->header('Location'), qr/Invalid...Key/, 'No error due to invalid key');
# config page for group 1 lists the new item
$res = $cb->(GET '/?rm=list_config&group_id=1',);
like($res->content, qr/ms_endpoint/, 'Config item ms_endpoint exists in group 1');
# adding service succeeds w/ correct key
$res = $cb->(POST '/', [
rm => 'insert_service',
'group_id' => 1,
'name' => 'ms',
'description' => 'Monitoring::Spooler',
'class' => 'MS',
'group_key' => 'test',
]);
ok($res->is_redirect, 'Redirect after create');
unlike($res->header('Location'), qr/Invalid...Key/, 'No error due to invalid key');
# config page for group 1 lists the new item
$res = $cb->(GET '/?rm=list_services&group_id=1',);
like($res->content, qr/App::Standby::Service::MS/, 'Config item ms_endpoint exists in group 1');
};
done_testing();