What can you monitor besides websites?
Websites are the obvious answer, but Oh Dear covers quite a bit more. We support four monitor types (HTTP, Ping, TCP, and AI), and each one opens up a different set of checks.
Websites and blogs
The bread and butter. Add any URL and we'll monitor uptime, performance, SSL certificates, broken links, mixed content, Lighthouse scores, DNS, domain expiration, and more.
Ideal for webshops, marketing sites, company pages, blogs, ... anything that speaks HTTP or HTTPS.
API endpoints
Any HTTP or HTTPS endpoint works. Send a GET, POST, PATCH, or PUT, attach headers or a payload, and we'll check it on every uptime run.
You can allow our IPs through your firewall or add custom authentication headers to reach protected endpoints. Great for testing login flows, order submissions, product lookups, or anything else you'd normally script.
JSON and XML response validation
Every uptime check can use the "look for string" option to assert that a specific snippet is present in the response body, whether that's HTML, JSON, XML, or plain text.
Combine that with a custom HTTP verb and payload, and you have a lightweight contract test running every minute.
S3 buckets and files
Both public and private S3 buckets work. For public ones, just add the bucket or file URL and we'll check that it returns an HTTP/200. For private buckets, add a custom authentication header so we can fetch the resource.
WebDAV and CardDAV
Both are built on top of HTTP, so we treat them like any other website. Add the base URL of your WebDAV or CardDAV endpoint and we'll check uptime, TLS certificates, and whatever else you've enabled.
WebSockets over HTTPS
The initial WebSocket handshake happens over HTTP(S). Add your WebSocket URL as if it were a regular website, including the port if it's running on something non-standard:
https://socket.example.com
That gives you certificate expiration tracking and uptime checks on the endpoint.
Servers and non-HTTP services
Not everything speaks HTTP. For raw server availability, use our other monitor types:
- Ping monitors use ICMP to check that a host is reachable. Handy for servers, routers, or appliances without a web interface.
- TCP monitors verify that a specific port (like
example.com:3306orsmtp.example.com:25) accepts connections. Perfect for databases, mail servers, SSH, and custom services.
Open ports monitoring also lets you track which ports are open or closed on a host and alert when that changes.
Full application health checks
Your site is typically a mix of application code, a database, queues, caches, and third-party services. Expose a /healthcheck endpoint in your application that reports on all of them, and let us hit it.
If your queue should always have fewer than 100 pending jobs, return an error when it doesn't. If a nightly import fails, flag it in the response. Oh Dear picks up the error and alerts you.
Two benefits:
- Your health logic lives in the code you already deploy, so it evolves with your app.
- You don't need to expose internal metrics publicly. Limit access to our IP addresses or send a shared secret header.
More in our application health docs.
Scheduled tasks and cron jobs
Every monitor can track scheduled tasks and cron jobs, on Linux, Windows, or anywhere else that can make an HTTP request.
Each task gets a unique ping URL. Your cron pings that URL when it finishes. If we don't hear from it on time, we let you know. More in the cron job monitoring docs.
Background workers and queues
The same ping system works great for long-running workers. Have your worker call the ping URL after it processes a batch, and you get a heartbeat on whatever your queue is actually doing.
AI-powered checks
Our AI monitor uses a prompt instead of a URL. Write "make sure the checkout page shows the EUR price" or "verify that the pricing table on our homepage is still visible" and we'll use AI to verify it for you, no brittle selectors required.
Still missing something you'd like to monitor? Tell us, we're always adding new checks.
