Contents

Monitors

A monitor answers one question: is this thing up? e-volv Observer has two kinds. An uptime check calls a URL on a schedule. A heartbeat waits for your job to call it. Both live under Observer, Monitors, and a workspace can have up to 20.

Uptime checks

Give it a URL, a method (GET, HEAD or POST), the status codes that count as up, and optionally a word the response body must contain. It runs every minute, 5 minutes, 15 minutes or hour, with a timeout between 1 and 30 seconds. Redirects are not followed, so a 301 is a 301.

A monitor goes down after two failed checks in a row (you can change the number) and comes back on the first success. Every check keeps its status and response time for 30 days, which is where the response-time chart and the 24-hour and 30-day uptime figures come from.

Checks only reach the public internet. Private, link-local and cloud metadata addresses are refused when you save the monitor, and checked again when each request connects, so a DNS name cannot be pointed inside later.

Heartbeats

A heartbeat gives you a ping URL. Call it at the end of every run, and set how often a run happens plus a grace period. If no ping arrives in that window, the monitor goes down. The URL needs no key: the token in it is the secret, so treat it like one.

bash
# at the end of the job
curl -fsS https://api.e-volv.io/api/public/v1/observer/heartbeat/<token>

# optional: mark the start, or report a failure straight away
curl -fsS https://api.e-volv.io/api/public/v1/observer/heartbeat/<token>/start
curl -fsS https://api.e-volv.io/api/public/v1/observer/heartbeat/<token>/fail

A cron line that pings only when the job succeeds:

bash
0 3 * * * /usr/local/bin/backup.sh && curl -fsS https://api.e-volv.io/api/public/v1/observer/heartbeat/<token>

/fail marks the monitor down at once. /start is recorded but changes nothing. The URL takes GET or POST.

Public status page

Open Monitors, Public status page to publish up to 20 monitors at e-volv.io/status/<slug>. Choose a public label for each monitor, whether to show response time, and post plain-text updates. The public view includes 30 daily uptime percentages and 14 days of updates. It never exposes monitor URLs, heartbeat tokens, ids or workspace ids. Unpublished and unknown pages return 404. A lapsed Observer subscription unpublishes the page.

Alerts

Create an alert rule of the kind An Observer monitor goes down or recovers under Settings, Alerts, and choose its channels. A down event and its recovery share one key, so a PagerDuty incident opens when the monitor goes down and resolves when it comes back. With no rule, the alert still appears in the app. See alert channels.

We use only essential cookies to keep you signed in and the app working. With your consent we may enable optional analytics to help us improve e-volv. See our Privacy Policy and Terms.