Methodology

Every outage in this dataset is a recorded event in our database — we mark a site as experiencing an incident when our automated HTTPS probe fails for two consecutive checks OR when crowd-sourced user reports spike to 5x or more above a 7-day rolling baseline. Total downtime is computed from incident start time to resolution time. Sites currently ongoing at report-generation are calculated up to the rendering moment. For the full methodology, see our methodology page.

Our HTTP probes run from a single US-based data center region (do-nyc3), so regional outages outside our checker location may be under-represented. Crowd-sourced reports help bridge that gap, but the data here skews toward outages observable from North America.

Outages by category — 90-day rolling

Cumulative downtime per service category, across all sites in our catalog.

Top 25 services by total downtime — 90 days

Sites with the most cumulative downtime over the past 90 days, summed across every detected outage.

Top 25 services by incident count — 90 days

Sites with the highest number of distinct outage events in the last 90 days. Many short incidents often signal infrastructure stability issues even when total downtime is moderate.

• Con Edison 298 incidents
• Petfinder 285 incidents
• Tencent Video 236 incidents
• Kuaishou 147 incidents
• Weibo 143 incidents
• Youku 137 incidents
• Cars.com 122 incidents
• VS Code 100 incidents
• Neeva 97 incidents
• Hilton 91 incidents
• Citation Machine 85 incidents
• Human Benchmark 76 incidents
• ePubPub 68 incidents
• Feishu 65 incidents
• KissAnime 52 incidents
• Jupiter 50 incidents
• Spotify 47 incidents
• iQIYI 46 incidents
• Allegro 45 incidents
• Framework 38 incidents
• LiveJournal 35 incidents
• OpenOffice 34 incidents
• JS Bin 33 incidents
• IndiGo 33 incidents
• Airtel 31 incidents

Citation

If you reference this data in a blog post, paper, or news article, please cite as:

Is it down checker Editorial Team. (2026).
"State of Website Uptime — 90-Day Rolling Reliability Report."
Is it down checker. https://isitdownchecker.com/research

Want a deeper cut of the data? Per-month reports going back to 2025 are at /reports; individual incident timelines are at /outages; and full methodology is documented at /methodology.