Google Cloud Ping Test

A Google Cloud ping test measures the round-trip latency between your browser and a Google Cloud public endpoint in each region. Unlike ICMP ping (which browsers can't send), it issues HTTPS HEAD requests so it works without local tools or a GCP project. Lower numbers point to the Google Cloud region that will give your users the lowest latency for Cloud Run, GKE, Cloud SQL, and similar workloads.

regionping pings the regional Cloud Run API endpoint (https://{region}-run.googleapis.com) for each Google Cloud region — the same hostname Cloud Run, Cloud Functions, and several other Google APIs use, so the latency you measure here closely tracks what your future workloads will see from the same network path. The browser sends one warmup HEAD request, then five timed HEAD samples, drops the high and low, and reports the median of three. Up to 16 regions are measured in parallel.

Your browser sends one warmup HEAD request per region to prime DNS, TCP, and TLS, then issues five timed HEAD requests. The highest and lowest samples are dropped and the median of the remaining three is shown. Up to 16 regions are measured in parallel.

regionping runs inside a browser, which cannot send ICMP packets. Every sample is an HTTPS HEAD request, so the measured time includes TCP and TLS overhead. Expect regionping numbers to sit roughly 10–30 ms above ICMP ping from the same machine. The ordering between regions is still faithful, which is what matters when choosing one.

AWS (32 regions), Google Cloud (41 regions), Azure (40 regions), Oracle Cloud (37 regions), DigitalOcean (10 regions), IBM Cloud (12 regions), Alibaba Cloud (29 regions), Linode (21 regions), OVHcloud (8 regions), Vultr (10 regions), Hetzner (3 regions), Huawei Cloud (26 regions), Exoscale (7 regions), Scaleway (4 regions), Gcore (3 regions), and Contabo (3 regions). 286 public regions in total.

Green (under 80 ms) is what you want for interactive workloads — API calls, real-time messaging, game servers. Yellow (80–149 ms) is acceptable for most web apps but noticeable in chatty request patterns. Red (150 ms and above) signals a region that is likely far from your network path; usable for batch and background jobs but a poor choice for anything user-interactive.

Most common causes, in roughly decreasing order of likelihood: a corporate firewall or enterprise proxy blocking the provider domain, an active VPN routing the request through a path that drops it, ISP-level blocks on cloud object-storage hostnames, the provider not yet deploying (or having deprecated) the public endpoint in that region, or a browser extension such as an ad blocker or privacy tool intercepting the request. Failures are surfaced explicitly instead of hidden so you can cross-check from a different network.