This is a framework agnostic cron jobs scheduler that can be easily integrated with your project or run as a standalone command scheduler. The idea was originally inspired by the Laravel Task Scheduling.
The recommended way is to install the php-cron-scheduler is through Composer. Please refer to Getting Started on how to download and install Composer.
After you have downloaded/installed Composer, run
php composer.phar require peppeocchi/php-cron-scheduler
or add the package to your composer.json
{
"require": {
"peppeocchi/php-cron-scheduler": "3.*"
}
}
Scheduler V3 requires php >= 7.1, please use the v2 branch for php versions < 7.1.
Create a scheduler.php
file in the root your project with the following content.
<?php require_once __DIR__.'/vendor/autoload.php';
use GO\Scheduler;
// Create a new scheduler
$scheduler = new Scheduler();
// ... configure the scheduled jobs (see below) ...
// Let the scheduler execute jobs which are due.
$scheduler->run();
Then add a new entry to your crontab to run scheduler.php
every minute.
* * * * * path/to/phpbin path/to/scheduler.php 1>> /dev/null 2>&1
That's it! Your scheduler is up and running, now you can add your jobs without worring anymore about the crontab.
By default all your jobs will try to run in background.
PHP scripts and raw commands will run in background by default, while functions will always run in foreground.
You can force a command to run in foreground by calling the inForeground()
method.
Jobs that have to send the output to email, will run foreground.
$scheduler->php('path/to/my/script.php');
The php
method accepts 4 arguments:
- The path to your php script
- The PHP binary to use
- Arguments to be passed to the script (NOTE: You need to have register_argc_argv enable in your php.ini for this to work (ref). Don't worry it's enabled by default, so unlessy you've intentionally disabled it or your host has it disabled by default, you can ignore it.)
- Identifier
$scheduler->php(
'path/to/my/script.php', // The script to execute
'path/to/my/custom/bin/php', // The PHP bin
[
'-c' => 'ignore',
'--merge' => null,
],
'myCustomIdentifier'
);
$scheduler->raw('ps aux | grep httpd');
The raw
method accepts 3 arguments:
- Your command
- Arguments to be passed to the command
- Identifier
$scheduler->raw(
'mycommand | myOtherCommand',
[
'-v' => '6',
'--silent' => null,
],
'myCustomIdentifier'
);
$scheduler->call(function () {
return true;
});
The call
method accepts 3 arguments:
- Your function
- Arguments to be passed to the function
- Identifier
$scheduler->call(
function ($args) {
return $args['user'];
},
[
'user' => $user,
],
'myCustomIdentifier'
);
There are a few methods to help you set the execution time of your schedules. If you don't call any of this method, the job will run every minute (* * * * *).
at
- This method accepts any expression supported by dragonmantank/cron-expression$scheduler->php('script.php')->at('* * * * *');
everyMinute
- Run every minute. You can optionally pass a$minute
to specify the job runs every$minute
minutes.$scheduler->php('script.php')->everyMinute(); $scheduler->php('script.php')->everyMinute(5);
hourly
- Run once per hour. You can optionally pass the$minute
you want to run, by default it will run every hour at minute '00'.$scheduler->php('script.php')->hourly(); $scheduler->php('script.php')->hourly(53);
daily
- Run once per day. You can optionally pass$hour
and$minute
to have more granular control (or a stringhour:minute
)$scheduler->php('script.php')->daily(); $scheduler->php('script.php')->daily(22, 03); $scheduler->php('script.php')->daily('22:03');
There are additional helpers for weekdays (all accepting optionals hour and minute - defaulted at 00:00)
sunday
monday
tuesday
wednesday
thursday
friday
saturday
$scheduler->php('script.php')->saturday();
$scheduler->php('script.php')->friday(18);
$scheduler->php('script.php')->sunday(12, 30);
And additional helpers for months (all accepting optionals day, hour and minute - defaulted to the 1st of the month at 00:00)
january
february
march
april
may
june
july
august
september
october
november
december
$scheduler->php('script.php')->january();
$scheduler->php('script.php')->december(25);
$scheduler->php('script.php')->august(15, 20, 30);
You can also specify a date
for when the job should run.
The date can be specified as string or as instance of DateTime
. In both cases you can specify the date only (e.g. 2018-01-01) or the time as well (e.g. 2018-01-01 10:30), if you don't specify the time it will run at 00:00 on that date.
If you're providing a date in a "non standard" format, it is strongly adviced to pass an instance of DateTime
. If you're using createFromFormat
without specifying a time, and you want to default it to 00:00, just make sure to add a !
to the date format, otherwise the time would be the current time. Read more
$scheduler->php('script.php')->date('2018-01-01 12:20');
$scheduler->php('script.php')->date(new DateTime('2018-01-01'));
$scheduler->php('script.php')->date(DateTime::createFromFormat('!d/m Y', '01/01 2018'));
You can define one or multiple files where you want the output of your script/command/function execution to be sent to.
$scheduler->php('script.php')->output([
'my_file1.log', 'my_file2.log'
]);
// The scheduler catches both stdout and function return and send
// those values to the output file
$scheduler->call(function () {
echo "Hello";
return " world!";
})->output('my_file.log');
You can define one or multiple email addresses where you want the output of your script/command/function execution to be sent to. In order for the email to be sent, the output of the job needs to be sent first to a file. In fact, the files will be attached to your email address. In order for this to work, you need to install swiftmailer/swiftmailer
$scheduler->php('script.php')->output([
// If you specify multiple files, both will be attached to the email
'my_file1.log', 'my_file2.log'
])->email([
'[email protected]' => 'My custom name',
'[email protected]'
]);
You can optionally customize the Swift_Mailer
instance with a custom Swift_Transport
.
You can configure:
subject
- The subject of the email sentfrom
- The email address set as senderbody
- The body of the emailtransport
- The transport to use. For example if you want to use your gmail account or any other SMTP account. The value should be an instance ofSwift_Tranport
ignore_empty_output
- If this is set totrue
, jobs that return no output won't fire any email.
The configuration can be set "globally" for all the scheduler commands, when creating the scheduler.
$scheduler = new Scheduler([
'email' => [
'subject' => 'Visitors count',
'from' => '[email protected]',
'body' => 'This is the daily visitors count',
'transport' => Swift_SmtpTransport::newInstance('smtp.gmail.com', 465, 'ssl')
->setUsername('username')
->setPassword('password'),
'ignore_empty_output' => false,
]
]);
Or can be set on a job per job basis.
$scheduler = new Scheduler();
$scheduler->php('myscript.php')->configure([
'email' => [
'subject' => 'Visitors count',
]
]);
$scheduler->php('my_other_script.php')->configure([
'email' => [
'subject' => 'Page views count',
]
]);
Sometimes you might want to execute a schedule not only when the execution is due, but also depending on some other condition.
You can delegate the execution of a cronjob to a truthful test with the method when
.
$scheduler->php('script.php')->when(function () {
// The job will run (if due) only when
// this function returns true
return true;
});
The jobs that are due to run are being ordered by their execution: jobs that can run in background will be executed first.
To prevent the execution of a schedule while the previous execution is still in progress, use the method onlyOne
. To avoid overlapping, the Scheduler needs to create lock files.
By default it will be used the directory path used for temporary files.
You can specify a custom directory path globally, when creating a new Scheduler instance.
$scheduler = new Scheduler([
'tempDir' => 'path/to/my/tmp/dir'
]);
$scheduler->php('script.php')->onlyOne();
Or you can define the directory path on a job per job basis.
$scheduler = new Scheduler();
// This will use the default directory path
$scheduler->php('script.php')->onlyOne();
$scheduler->php('script.php')->onlyOne('path/to/my/tmp/dir');
$scheduler->php('other_script.php')->onlyOne('path/to/my/other/tmp/dir');
In some cases you might want to run the job also if it's overlapping.
For example if the last execution was more that 5 minutes ago.
You can pass a function as a second parameter, the last execution time will be injected.
The job will not run until this function returns false
. If it returns true
, the job will run if overlapping.
$scheduler->php('script.php')->onlyOne(null, function ($lastExecutionTime) {
return (time() - $lastExecutionTime) > (60 * 5);
});
In some cases you might want to run some code, if the job is due to run, before it's being executed.
For example you might want to add a log entry, ping a url or anything else.
To do so, you can call the before
like the example below.
// $logger here is your own implementation
$scheduler->php('script.php')->before(function () use ($logger) {
$logger->info("script.php started at " . time());
});
Sometime you might wish to do something after a job runs. The then
methods provides you the flexibility to do anything you want after the job execution. The output of the job will be injected to this function.
For example you might want to add an entry to you logs, ping a url etc...
By default, the job will be forced to run in foreground (because the output is injected to the function), if you don't need the output, you can pass true
as a second parameter to allow the execution in background (in this case $output
will be empty).
// $logger and $messenger here are your own implementation
$scheduler->php('script.php')->then(function ($output) use ($logger, $messenger) {
$logger->info($output);
$messenger->ping('myurl.com', $output);
});
$scheduler->php('script.php')->then(function ($output) use ($logger) {
$logger->info('Job executed!');
}, true);
// $logger here is your own implementation
$scheduler->php('script.php')
->before(function () use ($logger) {
$logger->info("script.php started at " . time());
})
->then(function ($output) use ($logger) {
$logger->info("script.php completed at " . time(), [
'output' => $output,
]);
});
In some cases you might need to run the scheduler multiple times in the same script. Although this is not a common case, the following methods will allow you to re-use the same instance of the scheduler.
# some code
$scheduler->run();
# ...
// Reset the scheduler after a previous run
$scheduler->resetRun()
->run(); // now we can run it again
Another handy method if you are re-using the same instance of the scheduler with different jobs (e.g. job coming from an external source - db, file ...) on every run, is to clear the current scheduled jobs.
$scheduler->clearJobs();
$jobsFromDb = $db->query(/*...*/);
foreach ($jobsFromDb as $job) {
$scheduler->php($job->script)->at($job->schedule);
}
$scheduler->resetRun()
->run();
When running the scheduler you might pass an DateTime
to fake the scheduler run time.
The resons for this feature are described here;
// ...
$fakeRunTime = new DateTime('2017-09-13 00:00:00');
$scheduler->run($fakeRunTime);
If some job fails, you can access list of failed jobs and reasons for failures.
// get all failed jobs and select first
$failedJob = $scheduler->getFailedJobs()[0];
// exception that occurred during job
$exception = $failedJob->getException();
// job that failed
$job = $failedJob->getJob();