Skip to content
Alerts & on-call

Getting woken up well: the art of good alerting

Monitoring is easy; alerting is the hard part. Alert on everything and you'll start ignoring alerts. Alert on nothing and you find out from a customer. These are the levers WebWatch gives you, and when to pull each one.

The six levers

Each of these controls one thing: whether a problem interrupts you now, later, or not at all. Most sites only need two or three of them.

Confirmation windows

"Only alert me if it's down more than 5 minutes." The internet has hiccups: a routing blip, a server restart, a 30-second deploy. A confirmation window means an incident must persist before anyone is told, filtering out noise that would have resolved itself anyway.

Rule of thumb: a personal blog can wait 10 minutes; a checkout page gets 0 to 3.

Daily digests

Not every site deserves the power to interrupt your evening. Digest mode holds notifications and sends one morning email summarising outages, slow spells, and certificate warnings from the last 24 hours. Perfect for side projects and staging environments.

Real world: the difference between a newspaper and someone shouting headlines through your letterbox at 3am.

Acknowledge & escalate

Every downtime email has a one-tap Acknowledge button. Tap it and the system knows a human is handling it. Ignore it (because you're asleep, driving, or on a beach) and after a set number of minutes your backup contact gets paged automatically.

For teams: outages stop depending on one person's ringtone.

Performance budgets

Sites rarely fall over out of nowhere; they get slow first. A performance budget ("alert me if responses exceed 2 seconds") catches a struggling database, a filling disk, or a traffic spike while the site is still technically up.

Real world: the oil warning light, not the breakdown on the hard shoulder.

Webhooks

When a site changes state, WebWatch can POST a signed JSON payload to any URL you own. Pipe alerts into Slack or Discord, trigger a server restart script, flip a maintenance page on, or feed your own dashboard. Each delivery is HMAC-signed so you can verify it's really us.

For developers: alerts become events your own code can react to.

Public status pages

Turn your uptime into marketing. A public status page shows live state, 30-day uptime, and incident history for the sites you choose, at a link like web-watch.uk/status/your-name. During an outage it answers "is it just me?" so your inbox doesn't have to.

See it live: our demo status page tracks major websites in real time.

A setup that lets you sleep

A sensible default setup for a production site

Instant email alerts with a 3 to 5 minute confirmation window, SSL warnings at 30 days, a performance budget of 2 seconds, and a backup contact escalated after 10 unacknowledged minutes. For everything else (blogs, portfolios, staging) switch to the daily digest and sleep well.

Picking your numbers

Performance budgets make much more sense once you know what a normal response time looks like; the timing numbers are explained in the performance guide. And the statuses that actually trigger these alerts (DOWN, TIMEOUT, SSL_ISSUE and the rest) are decoded in the uptime guide.

Now try it on your own site

Run a one-off check or save a site to track uptime and SEO over time. The free plan covers 10 sites, no card needed.