Send real ICMP echo requests from a neutral host to any server. Runs 5 packets and reports min, avg, max, and packet loss.
Enter a hostname or IP to measure latency
This tool sends real ICMP echo requests - the same mechanism as the ping command on your own machine - from a neutral server in Vienna. Each row is one packet: the time shown is the round-trip time for that packet to reach the target and return.
The probe runs from a dedicated host rather than from a CDN edge node, which matters because ICMP from large cloud and CDN networks is frequently rate-limited or dropped outright. A neutral origin gives results closer to what a normal client would see.
Round-trip time is network latency only - it does not include TLS handshakes, server processing, or page rendering. That makes it a clean measure of the path between the probe and the host, and a poor measure of how fast a website feels. For that, use HTTP Timing, which breaks the request into DNS, TCP, TLS, and time-to-first-byte phases.
Under 30 ms - Excellent. Likely same-region or a nearby server.
30-100 ms - Good. Typical for intercontinental connections.
100-300 ms - Noticeable. Distant server or a congested path.
Over 300 ms - Poor. Satellite link, heavy congestion, or a badly routed path.
Timeout - No reply within the timeout. This does not mean the host is down: many servers and firewalls drop ICMP by design while serving HTTP normally. Check reachability with Is It Down or HTTP Timing before concluding anything from ICMP silence.
Loss shows how many of the 5 packets received no reply. Consistent partial loss usually means rate-limiting rather than a broken path - many hosts answer the first few pings and then throttle. Total loss almost always means ICMP is filtered, not that the host is unreachable.