How We Detect Outages

Community-driven reports

DownStatus combines multiple signals: official status feeds where available, user reports submitted on service pages, synthetic checks, and public discussion spikes. Official incidents carry the most weight; community reports help us detect regional or early problems faster.

Website checks

Every 20 minutes our server requests the website of each service and records whether it answers and how fast. Any HTTP response below 500 counts as up. Under 300 ms is fast, under 800 ms normal, under 2 seconds slow, and no answer within 3 seconds is a timeout.

Many large sites block requests from cloud servers. So when our check fails, we test the site again from North America, Europe and Asia. If most regions get a response, the service is up and only our server was blocked, and the service page shows it as reachable.

A few sites refuse automated checks from every region for days. For those we hide the website check and leave them out of the reliability rankings, and their status relies on user reports and, where available, the official status page.

Report types

Users can specify what kind of problem they are experiencing: website not loading, app issues, login problems, payment failures, streaming errors, or a general outage. This breakdown lets us show not just that a service is down, but which part is affected.

Outage detection threshold

We count user reports for each service over the last hour. 5 or more reports set the status to Degraded, 20 or more to Partial outage, and 50 or more to Major outage. If the service has an official status page, a major or critical incident there marks the service as down, and a minor incident marks it as degraded.

The same thresholds apply to every service. Our website checks and Reddit mentions do not change the status on their own: they appear as separate signals on the service page and count toward the confidence score.

Status levels

• Operational — report volume is within normal range. No significant issue detected.
• Degraded — elevated report volume. Some users may be experiencing issues.
• Outage — report volume is significantly above baseline. A widespread incident is likely in progress.

Spam and abuse prevention

To maintain signal quality we apply several layers of protection:

• Rate limiting — each IP address is limited to a small number of reports per service within a sliding time window.
• Duplicate suppression — a second report from the same IP for the same service within 5 minutes is silently discarded.
• Honeypot fields — hidden form fields invisible to real users but filled by bots; such submissions are discarded.
• IP hashing — IP addresses are never stored in plain text. They are one-way hashed before storage, so no personal data is retained.

The 24-hour chart

Each service page displays a bar chart of report volume over the past 24 hours, broken into hourly buckets. This gives a clear visual picture of when problems started, how severe they were, and whether the situation is improving or worsening.

Outage history

Confirmed outage events are stored and displayed in the Outage History section of each service page. Each entry shows the start time, end time (once resolved), and duration.

Independence

DownStatus is fully independent. We are not affiliated with, sponsored by, or endorsed by any of the services listed on this platform. We combine what users report with our own website checks and, where available, the company's official status page, so a problem can show up here even when the official page says everything is fine.

Questions?

If you have questions about our methodology or want to suggest improvements, visit our Contact page.