What it is for
The state of each watched URL right now, and the check history. The record screen says what to watch; this one says what is happening.
What is on it
- The state: up, down, or never checked.
- The last check: when it was, and what it answered.
- The response time, with history. This is where rising latency shows up before the outage.
- Check now: forces a check off schedule, to confirm it came back without waiting for the next cycle.
At a five-minute interval, one URL produces 288 checks a day and 105 thousand a year. The default retention is shorter than everything else on purpose: nobody asks about the latency of a Tuesday two years ago. It is adjusted under Settings, in the check purge policy.
When the URL requires a certificate
Some services (Brazilian tax authority web services, for example) require a client certificate and return 403 to anyone arriving without one. A plain GET cannot reach those, which is why the question asked of each URL is picked: the "NF-e authorizer status" one opens mutual TLS and reads the answer from inside the envelope — the HTTP code says 200 even when the service is down.
The certificate and its password live on the watched URL itself, in a block that only appears for URLs whose question asks for them. Whatever is on the part's card, under Parts and records, is the default for all of them: one certificate is uploaded once, and whoever issues for two companies puts the second one on the URL that needs it.
Each question has a minimum interval, and it belongs to the
question and not to this screen. The tax one is three minutes: asking for status
too often earns rejection 656, which blocks the company for an hour
for everything, including issuing invoices. A badly tuned
monitor stops the business from invoicing.