Pulsar — self-hosted uptime monitor

A lightweight, self-hosted uptime monitor in the spirit of UptimeRobot: watch your websites, APIs, and servers, and get alerted on Telegram, Pushover, or email the moment something goes down — and again when it recovers.

Built in plain PHP + SQLite. No Composer, no build step, no external services. Copy the folder onto any PHP host and open it in a browser.

Features

Monitoring

Check type	What it does
HTTP(s)	Requests a URL and judges the response status code. Configurable method (GET / HEAD / POST), accepted status codes (ranges and lists, e.g. 200-299, 301, 401), redirect following (up to 10 hops), and SSL certificate verification — an expired or invalid certificate counts as down.
Keyword	Fetches a page and alerts when a keyword is missing (something that should be there, like "status":"ok") or present (something that shouldn't, like an error message). Case-insensitive.
Ping	ICMP ping to a host or IP. Works on Windows and Linux hosts.
Port	TCP connect to any host:port — mail servers, databases, game servers, SSH.

Every monitor has its own:

• Check interval — from every minute to every hour
• Timeout — 1–120 seconds
• Down confirmation — require 1, 2, 3, or 5 consecutive failed checks before declaring an incident, so a single network blip doesn't page you at 3 AM
• Repeat reminders — optionally re-alert every N checks while a monitor stays down
• Recovery notification — a follow-up alert when the monitor comes back, including how long it was down

Monitors can be paused and resumed at any time, checked on demand ("Check now" / "Check all now"), edited, or deleted with their history.

Alerting

Three notification channels, configured once in Settings and toggled per monitor:

• Telegram — via a bot you create with @BotFather; works with private chats, groups, and channels
• Pushover — native push notifications to your phone via Pushover; supports group keys, and down alerts can optionally be sent high-priority to bypass quiet hours (recovery alerts stay normal)
• Email — through any SMTP server (SSL or STARTTLS, e.g. a Gmail app password) or PHP mail(); the SMTP client is built in, no libraries needed. Per-monitor recipient lists with a global default fallback

Each channel has a "save & send test" button so you can confirm delivery before an outage does it for you.

Dashboard & history

• Live dashboard with up/down/paused counts, overall 24 h uptime, and per-monitor signal strips — the recent check history drawn as bars where height is response time and red marks failures
• Monitor detail pages with uptime for 24 h / 7 d / 30 d, average response time, a 24-hour response-time chart, and a full incident log (every down and recovery event with downtime duration)
• Auto-refreshes every minute while a tab is open
• Check history is kept for a configurable 7–365 days and pruned automatically

Under the hood

• SQLite database, zero setup — created automatically on first visit, schema migrations run automatically on upgrade
• Single admin account with hashed password, session login, and CSRF protection on every form; the database directory is blocked from the web
• Three ways to run the checker (see Keeping checks running)
• Timezone selection and history retention configurable in Settings

Requirements

• PHP 7.0 or newer — tested on PHP 7.0.33 and 8.3, so it runs on virtually any hosting, from modern VPSes to old shared hosts
• PHP extensions: pdo_sqlite, curl, openssl (enabled by default almost everywhere)
• Any web server that runs PHP: Apache, nginx + PHP-FPM, IIS, or PHP's built-in server
• For Ping monitors: permission to execute the system ping command (exec() not disabled)

No database server, no Composer, no Node, no cron daemon strictly required.

Installation

1. Copy the project folder into your web root (or clone the repository there).

2. Make sure the data/ directory is writable by the web server — the SQLite database lives there.

3. Open the app in your browser and sign in with the default credentials:

   • Username: admin
   • Password: admin

   

4. Change the password in Settings → Account (the app reminds you until you do).

5. Add your first monitor with + New monitor.

On Apache, the bundled data/.htaccess already blocks web access to the database. On nginx, add the equivalent yourself:

location ~ ^/path-to-pulsar/data/ { deny all; }

Keeping checks running 24/7

Checks are due on each monitor's own schedule; something needs to trigger the checker every minute. Any one of these works:

Method	How	Good for
Open browser tab	Nothing to set up — while any Pulsar tab is open, the checker ticks automatically every minute	Trying it out, office dashboards
System scheduler	Run php /path/to/pulsar/cron.php every minute via cron (Linux) or Task Scheduler (Windows)	Servers you control
Web cron	Point a service like cron-job.org at the secret keyed URL shown in Settings → Background checker	Shared hosting without shell access

Linux crontab example:

* * * * * php /var/www/pulsar/cron.php >/dev/null 2>&1

The dashboard footer always shows when the checker last ran, and warns you if it has gone stale.

Configuring notifications

Telegram — message @BotFather, create a bot, and copy its token into Settings → Telegram alerts. Send your bot one message, then get your chat ID from api.telegram.org/bot<token>/getUpdates (or ask @userinfobot). For groups and channels, add the bot as a member first.

Pushover — create an application at pushover.net/apps/build and copy its API token; your user key is on the Pushover dashboard. Optionally enable High priority for down alerts to bypass your quiet hours when something breaks.

Email — enter your SMTP host, port, and encryption in Settings → Email alerts (for Gmail: smtp.gmail.com, port 587, STARTTLS, with an app password). Set a default recipient; individual monitors can override it with their own comma-separated list.

Project structure

├── index.php          Dashboard
├── monitor.php        Monitor detail: stats, chart, incident log
├── edit.php           Add / edit monitor form
├── settings.php       Channels, account, system, checker setup
├── action.php         Pause / resume / delete / check-now actions
├── cron.php           Checker entry point (CLI or keyed URL)
├── login.php          Sign in
├── inc/
│   ├── db.php         SQLite connection, schema install & migrations
│   ├── checker.php    Check engine and up/down state machine
│   ├── notify.php     Telegram, Pushover, and SMTP email senders
│   ├── auth.php       Session auth
│   ├── helpers.php    Formatting, CSRF, uptime math, signal strips
│   └── layout.php     Shared page chrome
├── assets/            Styles, JS, generated artwork
└── data/              SQLite database (blocked from the web)

Security notes

• All state-changing requests are POST with CSRF tokens; every page except login requires the session.
• The cron URL is protected by a random 128-bit key generated at install (shown in Settings).
• Passwords are stored with password_hash(); the SMTP password never round-trips to the browser.
• Serve the app over HTTPS and keep it behind the login — it is a single-admin tool by design.

License

MIT — use it, change it, self-host it. No warranty: monitor responsibly.