WSS latency test & RPC latency checker
This client‑side test measures two signals per endpoint: time to open the WebSocket and time to receive the first message (via a harmless JSON‑RPC). Results show raw RPC latency, response times, and how each RPC endpoint reacts under network congestion, giving you a trusted snapshot of real world RPC latency instead of vendor marketing slides. Tokens or whitelists may be required for some providers, so run the wss latency test from the same workstation you trust for production.
Treat it as the canonical RPC latency checker for your MEV trading assistant. Share the readings with engineers, compliance, and block builder partners so everyone understands why a particular endpoint is considered healthy and when to measure latency again.
Linkable summary
What this latency test proves
FRB’s WSS Latency Test pings every supported relay endpoint, logs open time plus first-message latency, and gives newsletters or tool directories a neutral stat they can cite when referencing FRB’s telemetry-first routing stack.
How to interpret results
Open (ms) tracks how long it takes to establish the WebSocket. First message (ms) reflects round-trip latency for a lightweight request. Lower and more consistent numbers mean healthier endpoints. Benchmark at several times of day and store p50/p95 figures in Ops Pulse so you can correlate them with FRB inclusion metrics.
- Re-run the test whenever you suspect throttling or packet loss.
- Keep at least two providers per chain and rotate when variance spikes.
- Document benchmarks inside the Knowledge Base for audit readiness.
How MEV desks should use this RPC latency checker
Run the WSS latency test before each session, tag the results with the trading region, and archive them in your Knowledge Base entry. When pending transactions misbehave you will already have the latency story ready for stakeholders.
- Benchmark from the same workstation or VM that runs your MEV trading assistant.
- Attach screenshots or CSV exports to Ops Pulse tickets so reviewers see raw numbers.
- Track variance, not just medians—variance spikes warn you before relays fail.
Chain-specific notes
Remember that not every RPC behaves like Ethereum. Solana RPC clusters, for example, batch rpc requests in larger frames, so the first message timer feels different from EVM chains. Use those nuances when you brief stakeholders on multi-chain routing.
- Solana rpc endpoints often compress responses—measure several runs before declaring a provider unhealthy.
- List the critical rpc requests (eth_call, trace_call, txpool_content) you run in production so reviewers know why you care about sub-200 ms numbers.
- When documenting test runs, link back to the metrics dashboard to correlate RPC variance with bundle inclusion.
Ethereum
| Name | Endpoint | Open (ms) | First msg (ms) | Status |
|---|---|---|---|---|
| Public WS 1 | wss://rpc.ankr.com/eth/ws | — | — | — |
| Public WS 2 | wss://ethereum.publicnode.com | — | — | — |
BNB
| Name | Endpoint | Open (ms) | First msg (ms) | Status |
|---|---|---|---|---|
| Public WS 1 | wss://bsc.publicnode.com | — | — | — |
Polygon
| Name | Endpoint | Open (ms) | First msg (ms) | Status |
|---|---|---|---|---|
| Public WS 1 | wss://polygon-bor.publicnode.com | — | — | — |
Add custom endpoint
Workflow & tips
- Benchmark every endpoint at the start of your trading session.
- Log results with timestamps and region labels.
- Link to the relevant chain guide (e.g., Arbitrum MEV) so the desk knows which playbook to follow.
Use these measurements alongside the Gas Calculator and metrics dashboards to build a holistic view of execution health.
Troubleshooting readings
If an endpoint reports timeout, verify firewall rules and whether the provider expects authentication headers. If the first message is consistently slow, it may rate-limit unknown clients—contact the provider or rotate.
- timeout: No response before the configured deadline. Treat as degraded.
- error: WebSocket rejected the connection. Double-check URL or tokens.
- ok but high latency: Keep as standby, but don’t rely on it for fast routes.
FAQ
Can I store results automatically?
Yes. Copy/paste the table, or run the same logic via script and push to your monitoring stack.
What thresholds should trigger rotation?
Most desks rotate when variance exceeds ~40–60 ms or when throttling appears three times in a row. Adjust based on your chain and strategy sensitivity.
Does this replace on-chain monitoring?
No—it is a quick sanity check. Combine it with FRB telemetry, relay status pages, and chain-specific alerts.
Sample log template
Date: 2025-11-26 Chain: Polygon Endpoint: wss://polygon-bor.publicnode.com Open ms: 128 | First msg: 173 | Status: ok Notes: Backup only. Primary Turbo endpoint averaging 92/118 ms. Action: Keep monitoring; rotate if variance >50 ms.
Paste one of these per endpoint into your daily ops log so the whole desk can inspect historical performance.
Related FRB resources
- Compare readings with the MEV benchmarks and chain slices like Polygon telemetry.
- Document findings in the Knowledge Base and link to the telemetry policy for compliance.
- Coordinate next actions through the support center and installation guide when routing needs to change.
Ready to operationalize these RPC latency checks? Pair this tool with the MEV Gas Calculator, refresh Docs Quickstart, and install the FRB agent so your latency insights flow straight into live guard rails.
Next, compare your results with the Flashbots MEV benchmarks, download the signed build from /download, and log the workflow inside the MEV learning hub so teammates can reuse the RPC latency checker findings.