Cloud Region Latency Testing
Pick the fastest region for your deployment by measuring real HTTPS response times from your location.
What You're Actually Measuring
When you open a provider page (like /aws), your browser sends real fetch() requests to endpoints in each region. Each endpoint is hit twice:
| Metric | What It Includes | Best For |
|---|---|---|
| Setup time | DNS lookup + TCP connection + TLS handshake + first HTTP round trip | Estimating cold-start latency (first user request) |
| Response time | HTTP round trip over an already-open connection | Estimating steady-state API call latency |
⚠ This is not ICMP ping. Network engineers use ICMP ping for raw network-layer round-trip time. Our numbers are higher because they include the full HTTPS stack, but they're also more representative of what your application actually experiences. Think of it as application-layer latency, not network-layer latency.
How to Pick a Region
- Open a provider page below. All regional endpoints start checking automatically.
- Wait for results. The main number (in bold) is your steady-state response time. The smaller "Setup" number is the cold-start overhead.
- The lowest response time = closest region. If two regions are within ~10ms, they're effectively equivalent, so pick based on other factors (pricing, compliance, services).
- Run checks at different times of day. Congestion and routing can shift results, especially for intercontinental paths.
Reading the Numbers
| Response Time | What It Means |
|---|---|
| < 50ms | Same region or nearby. Excellent for latency-sensitive apps |
| 50–150ms | Same continent, different city. Good for most workloads |
| 150–300ms | Cross-continent. Acceptable for APIs, noticeable for real-time |
| > 300ms | Opposite side of the world. Consider a CDN or multi-region |
Test by Provider
Each page checks endpoints across that provider's global regions:
Test by Geography
Compare all providers in a specific part of the world:
Pro Tips
- Bookmark it. This is a lightweight PWA. Install it on your phone or laptop and test connectivity from any network you're on.
- Use Custom Lists for internal infra. Have staging servers, internal APIs, or health-check endpoints? Build a custom list with those URLs and bookmark it.
- Compare to server probes. Below each endpoint, expand "Server Probes" to see response times from our servers in different regions. Useful for ruling out local network issues.
- The site loads ~200KB. No frameworks to download, no ads, no trackers. The entire app is pre-built as static files served from a CDN.