Tracks page views of your Laravel 5.x app for traffic monitoring.
To date, it has been tested with Laravel up to 5.5 and PHP 7.
IMPORTANT If you are upgrading from previous 0.x version please read the release notes.
This package is meant for simple request tracking and not for in-depth traffic analysis. Features:
- track page views
- track unique page views
- customizable whitelist rules
- url filter and crawler filter
- fetch data for reports
Each log record keeps track of the user session, user ip and date.
Install module:
composer require foothing/laravel-simple-pageviews
Add the service provider in config/app.php
:
'providers' => [
Foothing\Laravel\Visits\ServiceProvider::class,
]
Publish migration and configuration files:
php artisan vendor:publish provider="Foothing\Laravel\Visits\ServiceProvider --tag="config"
php artisan vendor:publish provider="Foothing\Laravel\Visits\ServiceProvider --tag="migrations"
Run the migration:
php artisan migrate
Finally, add the middleware in your app/Http/Kernel.php
:
protected $middleware = [
'Foothing\Laravel\Visits\Middlewares\CountPageView',
];
In config/visits.php
// Enable or disable tracking.
"enabled" => true,
// Add patterns to be blacklisted (ignored and not tracked).
// These patterns will apply if the UrlWhitelist rule is
// enabled in the chain.
"blacklist" => [
// i.e. '/^(admin|api|auth).*/'
],
// Rules chain.
"rules" => [
"Foothing\Laravel\Visits\Rules\Crawler",
"Foothing\Laravel\Visits\Rules\UrlWhitelist",
],
Rules are meant to filter requests that you don't want to track. Default ones will filter out crawlers (thanks to https://github.com/JayBizzle/Crawler-Detect) and blacklisted urls.
$manager = app()->make("Foothing\Laravel\Visits\Reports\ReportManager");
// Get visits with url and hits. This should be used to
// get an overview of the best performer urls.
$manager->getVisits();
[
{
"url": "foo/bar",
"hits": 129
},
{
"url": "baz",
"hits": 40
}
]
// Return int
$manager->countOverallVisits();
$manager->countOverallVisits('today', 'tomorrow', 'foo/bar');
// Return int
$manager->countUniqueVisits('currentWeek');
$manager->countUniqueVisits('currentWeek', null, 'foo/bar');
// Return a data collection meant for chart plotting.
$manager->getVisitsTrendDaily();
[
{
"day": "2016-10-08"
"hits": 129
},
{
"day": "2016-10-09"
"hits": 40
}
]
Each query method allows for date filtering and will accept up to 3 arguments.
- first argument can be a string preset, a string that
Carbon
can parse or aCarbon
date - second argument as the first one
- third argument can be an url
Only exception is the aggregate()
method, which will accept a limit
argument.
Examples:
// Triggers default filter
$manager->getVisits();
// String presets
$manager->getVisits('today');
$manager->getVisits('currentWeek');
$manager->getVisits('currentMonth');
$manager->getVisits('currentYear');
// Single day
$manager->getVisits(\Carbon::now());
// Period
$manager->getVisits(\Carbon::now(), Carbon::tomorrow());
This package has been tested in a moderate traffic website, like ~20k pageviews / day
which makes about 17k database records per day. The visits table will grow up pretty
quick and the database might suffer performance issues when it comes to execute an
insert or update
statement on a table which count hundred thousands (or millions)
rows.
For this reason an insert/update buffer has been added. Basically, each visit is tracked in a temporary table that is only used on write operations, while the report and read operations are performed on a separate table.
An artisan
command has been added to handle the periodic data dump from the
write table to the read table. A good practice might be dumping data each day.
TL;DR configure as follows in your app/Console/Kernel.php
(please refer to Laravel docs for scheduling info):
/**
* The Artisan commands provided by your application.
*
* @var array
*/
protected $commands = [
'Foothing\Laravel\Visits\Commands\DumpVisitsBuffer',
];
/**
* Define the application's command schedule.
*
* @param \Illuminate\Console\Scheduling\Schedule $schedule
* @return void
*/
protected function schedule(Schedule $schedule)
{
// Adjust this with your needs.
$schedule->command('visits:buffer')->dailyAt("00:00");
}
Performances have been tested in a ~5 million records database with good results. However, i recommend to partition your database tables if size grows nasty, i.e. 1 milion record per partition. Also, a good practice would be to tune your partition to be consistent with date periods (i.e. full year in single partition, single quarter in single partition, etc.) according to your traffic and report type.
MIT