The Edge tier is the production endpoint for GoldRush JSON-RPC. It runs from regional edge POPs in front of an analytics-driven multi-provider router, with archive depth, theDocumentation Index
Fetch the complete documentation index at: https://goldrush.dev/docs/llms.txt
Use this file to discover all available pages before exploring further.
debug_* and trace_* namespaces, and an SLA.
Endpoint
What you get
| Feature | Edge |
|---|---|
eth_*, net_*, web3_* methods | ✅ |
debug_* namespace | ✅ |
trace_* namespace | ✅ |
| Archive depth (full historical state) | ✅ |
| Sub-100ms p50 latency | ✅ |
| 99.9% uptime SLA | ✅ |
| Rate limit | Account-scoped, configurable |
| Multi-provider failover | ✅ |
How requests are routed
The Edge layer:- Authenticates the request, so invalid keys never hit the upstream.
- Caches safe-to-cache reads (e.g.
eth_getBlockByNumberfor finalized blocks) at the edge. - Routes the call to whichever upstream provider currently scores highest on cost and performance for that specific
(chain, method)route. Scores are continuously updated from rolling-window success rate, latency, error class, and cost analytics. - Re-routes on failure to the next-best provider, and feeds the failure back into the score so the routing table self-corrects in seconds.
Archive depth
Edge serves historical state and traces back to the chain’s earliest available archive block. Depth varies by chain. See debug and trace for the per-chain matrix.SLA and rate limits
- Uptime SLA: 99.9% per calendar month
- Rate limit: configurable per account; defaults to 1,000 RPS
- Burst: 2× the configured RPS for up to 60 seconds
debug_* and trace_*
Method list and per-chain availability.
Failover and SLA
How the Edge layer handles upstream failures.
Pricing
Per-method credit rates.
Migration
Side-by-side diffs from other providers.