dnspup probing
With your own dnspup customer API key, probe origins, third-party CDN access points and nodes from dnspup checkpoints across China, independent of node probes.
What it solves
Node-to-node probes misjudge when a node's own network is the problem, and probes from the control plane come from a single place. dnspup probing adds dnspup's checkpoints across every Chinese carrier as an independent source:
- Origin and aggregated edge probes: every checkpoint counts as one vote, merged with the node votes.
- Node availability probes: each node IP's ports are probed from carrier checkpoints on a schedule, with results in the node's monitoring log.
- Latency for smart scheduling: line groups use this data to pick nodes per line.
Turning it on
dnspup probing is an official, free module built into both editions; the probing quota comes from your own dnspup account.
- Open the customer API in dnspup and get a key and secret.
- Open System → dnspup probing, turn on Enable dnspup probing and enter the API address, key and secret; fill in the bound domain if the key is bound to a server domain, leave it empty for a fixed IP.
- Click Test connection and capabilities; save once it reports the connection is fine and shows this month's quota.
- Under Node availability probing, choose the interval and the checkpoint carriers and regions.
Main settings
| Setting | What it does |
|---|---|
| Key and secret | The secret is encrypted and shown as a fingerprint; leave empty to keep, enter a new one to rotate |
| Origin probe source order | Node probes, dnspup checkpoints and control plane probes are asked in order; when one source has too few votes (under 2, or a tie) the next fills in |
| Aggregated edge source order | The same, for third-party CDN access points |
| Node availability interval | Off, or every N minutes |
| Checkpoint carriers / regions | Which checkpoints probe your nodes |
Typical uses
- Origin overseas, nodes in China: use dnspup checkpoints as a second opinion so a node's own network trouble does not switch origins by mistake.
- Carrier-aware scheduling: once enabled, line groups' smart scheduling has latency data.
Limits and notes
- With quota exhausted, rate limited or the key rejected, probing falls back to nodes and the control plane.
- dnspup HTTP probes can only connect to the URL's own host: origins whose Host differs from the connect address, or that use a custom CA, are not sent to dnspup; third-party CDN access points are judged by a TCP connection to 443.
- Node availability: DNS for a node is paused only when the local probe and most dnspup checkpoints fail; dnspup failing alone only alerts.
- When unlicensed, disabled, missing a key or paused, nodes are judged by local probes only.