Checks
Every check has a name, a type, a target and an interval. A check is UP or DOWN; a new check is
UNKNOWN until its first run. Before an UP check turns DOWN the probe is repeated (2 times, 10 seconds apart by
default) — a one-second network blip is not an alert.
| Field | Default | Meaning |
|---|---|---|
name |
— | shown in the UI, alerts and the status page |
type |
— | HTTP, TCP, CERT or PUSH |
target |
— | HTTP: URL; TCP: host:port; CERT: host or host:port |
intervalSeconds |
60 | how often to check (PUSH: how often a signal is expected) |
timeoutMs |
10000 | how long to wait for an answer |
statusPage |
false | show on the public status page |
group |
— | group on the status page |
HTTP(S)
UP when the response code is at most maxStatus (default 399 — redirects are fine, errors are not).
| Field | Meaning |
|---|---|
expectedStatus |
exactly this code, e.g. 401 for a page behind a password |
maxStatus |
UP while the code is not above this; 499 — only 5xx and no answer are DOWN |
keyword |
the page must contain this text |
TCP
UP when the port accepts a connection within the timeout.
CERT
Checks the whole TLS chain the way a browser does. DOWN when the chain is not valid or the certificate ends in fewer
than certWarningDays days (default 14).
PUSH
Your job calls the check's push URL when it finishes. No call within intervalSeconds + graceSeconds — DOWN.
The URL is in the API answer (pushUrl) and on the check page.
# at the end of a backup script
curl -fsS https://status.example.com/push/<token>
# report a failure at once
curl -fsS "https://status.example.com/push/<token>?status=down&msg=disk%20full"