Getting started

Run the first check

Create an account. Signup is self-service with a work email and an activation link, and it needs no credit card. You can run real checks against your own services before you pick a plan.
Point a monitor at a URL. Pick http, paste the URL, and keep the defaults. The monitor probes from the regions in your plan, and the standard alert profile waits for a second region to confirm a failure before an incident opens. The quorum rule explains the details, including the monitor types that report directly.
Watch the first results. Per-region latency, per-IP results for dual-stack targets, and DNS time separated from total time arrive with the first checks.
Test an outage and the recovery. On an endpoint you control, briefly set the content assertion to a string the response does not contain. Check the failed result, restore the assertion, and confirm a fresh passing result. Whether this opens an incident depends on the check type and the alert profile.

Cover the essentials in the first hour

Add an ssl_cert monitor for the certificate you always forget, and a domain monitor for the registration layer (NS, DNSSEC, WHOIS expiry). The check types each explain what they verify.
Create your status page and pick the components worth showing. Publish it openly, or protect it with a password or viewer accounts from Sentinel.
Verify your phone number with a one-time code, so Perstat can reach you by phone. Personal SMS is included from Pulse, on-call SMS and calls from Sentinel.
Announce your first maintenance window before you deploy. Announced windows never count against your availability.

Where things live

The cockpit is the live view across everything. History carries the raw timeline, the curated outage record, and the configuration audit. Every number you see comes from the same availability record that your status page and your SLA reports use. One record, no drift.

Next: publish a status page to share the situation, or set up alerting and on-call for the people who respond.