Uptime monitoring

Uptime monitoring checks your pages around the clock and tells you when they stop responding. Every failure is re-checked from a second location before you're alerted, so a short network hiccup doesn't wake you up at night. You'll also see response times, a 90-day history and reminders before your SSL certificate or domain registration runs out.

You'll find it in the sidebar under Project → Uptime (open it).

How it works

A monitor checks one URL at a fixed interval. A check passes when the page answers in time with the expected HTTP status, and, if you set one, when the keyword rule matches.

• Location: the region your checks should come from. Pick a country, or Auto to let serp.ac choose. If the chosen location isn't available at check time, another location is used automatically. Your monitor always shows the location you configured.
• Check every: from 1 minute to 1 hour. The shortest interval depends on your plan.
• Timeout: how long to wait for an answer, from 5 to 30 seconds (default 15).
• Expected status: empty means any 2xx or 3xx answer counts as up. You can also list codes and classes, such as 200 or 200, 301 or 2xx. Redirects are followed, up to 5.
• Keyword: optional text that Must appear (for example a word from your footer) or Must not appear (for example "Database error") in the first 256 KB of the page. The search ignores upper and lower case.

When is a site "down"?

1. A check fails, for example because of a timeout, a DNS error, a 5xx status or a missing keyword.
2. About 45 seconds later, serp.ac re-checks from a different location.
3. The site only counts as down if the re-check fails too. The incident then starts at the time of the first failure, and you get one "down" alert.
4. As soon as a check passes again, the incident is closed and you get one "back up" alert that says how long the site was down.

If the re-check passes, it was a false alarm: you get no alert and it doesn't count as downtime.

Plan limits

Plan	Monitors	Shortest interval
Free trial	2	5 minutes
Starter	5	5 minutes
Pro	20	3 minutes
Agency	50	1 minute

The Links plan doesn't include uptime monitoring. Monitors count across all your projects. When your plan ends, checks pause and your monitors and history are kept. See Billing for plans.

Step by step

1. Open Project → Uptime and click Add monitor (shortcut).
2. Enter the URL. It's filled in with your project's homepage.
3. Optionally give it a Name, such as "Homepage" or "Checkout".
4. Pick a Location and how often to check (Check every).
5. Under Advanced, you can change the Timeout, the Expected status and a Keyword rule.
6. Click Add. The first check runs within about 30 seconds.

To change a monitor later, open it and click Edit monitor. Pause stops the checks without losing the history, and Delete removes the monitor and its history.

What the numbers mean

Metric	Meaning
Uptime · 24 h / 7 days / 30 days	The share of time the site was up. Only confirmed incidents count as downtime, and only time since you created the monitor counts.
Avg. response · 24 h	The average time until your server starts answering, over the successful checks of the last 24 hours. p95 is the time 95% of checks stayed under.
Response time	15-minute averages over the last 24 hours.
Last 90 days	One bar per day. Green means no downtime, yellow means some downtime (up to 5% of the day) and red means a major outage. Grey means there's no data, for example before the monitor existed. On a phone, the last 30 days are shown.
Incidents	Every confirmed outage, with its start, duration and cause, such as "Timed out", "HTTP 503" or "Keyword check failed".

SSL certificate and domain expiry

For HTTPS addresses, serp.ac checks the SSL certificate once a day and reads the domain's registration date (via RDAP) once a week, or daily when the expiry is close. You get an alert:

• SSL certificate: 14 and 3 days before it expires, when it has expired and when browsers don't trust it (self-signed, wrong host name or an incomplete chain);
• Domain: 30 and 7 days before the registration expires.

Some domain endings (TLDs) don't publish their expiry date. Then the domain card says it's not available, and you get no domain alert.

Alerts

Two events in notification settings control where uptime alerts go:

• Website down / back up (uptime monitors): by email and Telegram by default.
• SSL certificate and domain expiry: by email and Telegram by default.

Slack and Discord follow the Telegram setting by default. See Notifications.

FAQ

Why wasn't I alerted about a short outage?

Each failure is re-checked from another location about 45 seconds later, and you're only alerted when both checks fail. A blip that is over by the re-check doesn't trigger an alert and doesn't count as downtime. To catch shorter outages, use a shorter interval (Check every) if your plan allows it.

Do the checks really come from the country I picked?

Checks are run from data centers in the region of the country you picked whenever that's available. If it isn't available at check time, another location takes over automatically, so your monitor never misses a check. That's why the monitor shows your configured location, not the location of each single check.

Will the checks show up in my analytics or overload my server?

Checks are single GET requests with a user agent that contains serp.ac-uptime, so you can filter them out of your logs and analytics. They don't run JavaScript, so tools like Google Analytics don't count them. Checks of the same site are spread out over time instead of arriving in bursts.

Can I monitor a page behind a login or a non-public address?

No. Monitors check public pages. Private network addresses, IP addresses and pages that need a login can't be monitored. Monitor the login page itself, or a public health-check URL of your app.

What counts as a keyword failure?

With Must appear, the check fails when the text isn't in the first 256 KB of the page. With Must not appear, it fails when the text is there. Use it to catch pages that answer "200 OK" but show an error message.

Related guides

• Notifications
• Site audit
• Billing