Latency Testing and Auto-Select: url-test Groups Explained
Tired of switching nodes by hand? A url-test policy group re-tests its nodes on a schedule and automatically routes through the fastest — the key to a set-and-forget setup.
How the latency test works
The client makes one request through each node to a test URL (usually a 204 page on gstatic) and records the time. It measures real browse-through-the-node latency, not a bare ping.
Parameters that matter
- url: the test target — the default is fine
- interval: seconds between automatic tests; 300 is a good value
- tolerance: a new node must beat the current one by this margin to trigger a switch — 50 ms stops pointless flapping
When tests time out
A few nodes timing out is normal — provider nodes rotate. All of them timing out means check your own network first, then whether the plan expired. And remember: low latency is not high bandwidth; downloads also depend on the node's capacity.
Ready to get started?
Download Clash Verge Rev, FlClash or the mihomo core and be up in minutes.