Monitoring
Uptime monitors
Check a URL every 5 minutes, every minute or every 15 seconds, see its uptime and response time, and get alerted when it goes down.
An uptime monitor requests one URL from Logdash servers at a fixed interval and records every check: the status code, the response time and, when it fails, why. When two checks in a row fail, Logdash alerts the channels of the domain.
Logdash has two kinds of monitors. This page covers the monitors that check a URL. For a worker or another process that has no URL, use a heartbeat monitor: your process calls Logdash, and Logdash alerts you when the calls stop.
Add a monitor
Your domain gets its first monitor when you add it. To watch more addresses, such as your API or a health check endpoint:
- Open Uptime Open your domain, go to Uptime and press Add monitor.
- Enter the URL Keep We check a URL, enter the full URL, such as
https://api.example.com/health, and give the monitor a name. - Start monitoring Press Start monitoring. The first check runs within about a second, and the next ones follow the interval of your plan.
Point monitors at a URL that proves your app works, such as a health endpoint that touches the database. Many hosts answer every path with the same page, even when the app behind them is broken. When Logdash sees that, the monitor page warns you and offers /health, /api/health or /up if your host serves one of them.
How a check works
| What | How |
|---|---|
| Request | A GET with no extra headers and no body. |
| Up | Any status code from 200 to 399. Everything else is a failed check. |
| Timeout | 10 seconds. A slower answer fails with "Timed out after 10s". |
| Redirects | 301, 302, 303, 307 and 308 are followed, up to 5 in a row. |
| Response body | Read up to 5 MB. |
| Addresses | Every host, including each redirect target, has to resolve to a public IP address. Private and local addresses fail with "Hostname does not resolve to a public address". |
A check that got no answer records status code 0 and a reason, such as "Connection refused", "Connection reset" or "DNS lookup timed out". A check that got an error status records the start of the response body as its message, so you can see what your server said.
When most monitors get no answer in the same round, the problem is on the side of Logdash, not yours. Logdash drops those checks instead of recording them, so an outage of its own never alerts you.
Check interval
Monitors check every 5 minutes on Hobby, every minute on Builder and every 15 seconds on Pro. The interval follows the plan of the domain owner and applies to every monitor of the domain.
Status
Each monitor has a status, computed from its latest checks:
| Status | When |
|---|---|
| Down | The 2 latest checks both failed. |
| Degraded | Any of the last 10 checks failed, but not the 2 latest. |
| Up | None of the last 10 checks failed. |
| Unknown | The monitor has no checks yet. |
Alerts, status pages and badges all use these rules, so a single failed check never pages you. The Logdash dashboard is stricter: it shows a monitor as down from its first failed check, so you see trouble before anyone else does.
The Uptime page
The Uptime page of a domain shows how many monitors are down, the average response time and the uptime over the range you pick, 90 hours or 90 days. Below it, each monitor has a row with its latest checks.
Monitors of well-known services you depend on, such as Stripe, GitHub or OpenAI, are listed apart under Dependencies, so your own services stay on top.
The monitor page
Click a monitor to open its page. It shows the status and the latest checks, live, with the status code, response time and error message of each. Below that, Uptime history has one bar per hour or per day for the range you picked.
- Alerts lists the channels of the domain. Uncheck a channel to stop alerts from this monitor only. See Alerts.
- Edit, in the Monitor card, changes the name and the URL of the monitor.
- Get badge gives you a README badge with the uptime of the monitor. See README badges.
- The trash button of the Monitor card deletes the monitor with its uptime history and alerts. This cannot be undone.
There is no pause. To stop checking an address, delete its monitor.
Plan limits
The number of monitors counts across all the domains you own. Monitors in domains you were invited to count toward the plan of their owner.
| Limit | Hobby | Builder | Pro |
|---|---|---|---|
| Monitors, across all your domains | 5 | 20 | 50 |
| Check interval | 5 minutes | 1 minute | 15 seconds |
| Heartbeat monitors | No | No | Yes |
| Status pages | 1 | 5 | 15 |
| Custom status page domain | No | No | Yes |