Practice: Design an Uptime Monitor
Uptime monitoring service inspired by UptimeRobot and Pingdom: probes report check results, public status pages are cached and served from the edge, and outage alerts go out through a background queue.
Build this system yourself on the canvas below. Add components, connect them, and run traffic simulations to see where your design holds up and where it breaks.
Functional requirements
- Show a site's current status (up, down, degraded) and its uptime history on a public status page
- Accept check results from probes in different regions and decide a site is down only when several regions agree, to avoid false alarms
- Send an alert (email, SMS, webhook) when a site goes down, and again when it recovers
- Keep 90 days of check history so uptime percentages and incident timelines can be shown
- Let an owner configure what to check (URL, interval, expected response) and who gets alerts
- Keep working during an incident: a surge of viewers on the status page must not slow down result intake
Non-functional requirements
- Status page < 100ms p99, served from the CDN and status cache
- A down site is shown within 1 minute, and the alert is sent within 1 minute of the outage
- Handle 10K status-page requests per second at peak and 100 probe results per second
- No flapping alerts: a site must fail 2 or 3 checks in a row before it is called down
- The status page stays up even if the monitoring pipeline is down (99.99% availability)
- Scale to 10M monitored sites; history older than 90 days is dropped or rolled up
- Alerts are delivered at least once, with retries if the provider is down