Our 5-Signal Outage Detection System
IsDownNow monitors 300+ platforms, services, and cloud infrastructure around the clock using a distributed multi-signal consensus engine. Here is how our detection architecture works in real time.
Signal 1: real-time search trends Breakout Velocity
We monitor search query acceleration across real-time search trends for terms like “[service] down”, “[service] not working”, and “[service] error”. When query velocity breaks out beyond historical baselines, this provides an early, independent indicator of user-facing disruption.
Signal 2: Official Statuspage Ingestion
We poll official status pages (Atlassian Statuspage, incident.io, custom status feeds) across 100+ providers. When a provider flags degraded performance or major outages on their status dashboard, that evidence is factored into our consensus scoring.
Signal 3: Crowdsourced Incident Reports
Real users visiting IsDownNow submit symptom reports (e.g. login failed, server connection timed out, app crash). Volume and velocity anomalies are normalized against time-of-day traffic models.
Signal 4: Multi-Region Network Latency Probes
Synthetic probes measure round-trip latency, DNS lookup times, and HTTP status codes across distributed edge nodes, verifying whether service APIs are reachable or returning gateway errors.
Signal 5: Multi-Vector Consensus Scoring (0–100)
All independent signals are aggregated into a unified Outage Score:
- 0–24: Operational (Nominal) — Normal traffic, response times within normal baselines.
- 25–64: Degraded Performance — Elevated complaints or regional latency spikes.
- 65–100: Confirmed Outage — Multi-source verification confirms widespread service failure.
Update Frequency & Global Coverage
Our distributed crawler swarm updates telemetry every 3 to 4 minutes, 24 hours a day, 7 days a week. We track regional disruptions across North America, Europe, Asia-Pacific, Latin America, and India.
Data Accuracy & Community Feedback
While our consensus engine minimizes false positives, regional or carrier-specific issues can vary. If you believe a service score requires adjustment or would like to suggest a new service to monitor, contact our team.