ap-pool-02 is answering in 310 ms
9 accounts are bound to it. Proxy latency is added to every order those terminals place, and it is the one component of the copy path that is invisible in a trade’s own timings — it looks like broker slippage.
Proxies65 usable
Accounts bound79routed through a proxy
Slow1above the latency budget
Down1not answering
The proxy pool
Accounts per IP, and what each one costs in latency
| Label | Protocol | Endpoint | Region | Accounts | Latency | Health | |
|---|---|---|---|---|---|---|---|
| eu-pool-01 | SOCKS5 | 51.89.104.22:1080 | eu-west-1 | 18 | 42 ms | Healthy | |
| eu-pool-02 | SOCKS5 | 51.89.104.23:1080 | eu-west-1 | 16 | 38 ms | Healthy | |
| ap-pool-01 | SOCKS5 | 139.59.20.14:1080 | ap-south-1 | 22 | 71 ms | Healthy | |
| ap-pool-02 | SOCKS5 | 139.59.20.15:1080 | ap-south-1 | 9 | 310 ms | Slow | |
| us-pool-01 | HTTP | 167.71.40.9:8080 | us-east-1 | 14 | 58 ms | Healthy | |
| us-pool-02 | HTTP | 167.71.40.10:8080 | us-east-1 | 0 | — | Down |
Why the pool is managed
Three things that are not obvious from the table
- Accounts per IP is the figure that matters. Brokers notice when many accounts trade identically from one address, and what they do about it ranges from nothing to closing the accounts.
- Proxy latency is added to every order and is invisible in the trade’s own timings — it presents as broker slippage, which is why it is measured here separately rather than inferred from fills.
- A down proxy takes every account bound to it offline. It is not offered when assigning, and rebinding away from it is the fix rather than waiting.
- Proxy credentials are write-only, the same rule as everything else. There is no field showing the current one.