Reliability
All chains operational
Measured, not promised: synthetic JSON-RPC probes hit every upstream endpoint every five minutes, around the clock. A chain counts as up when at least one upstream answered — the same failover every relayed request gets. These numbers cover our full endpoint pool, including premium supply; free-tier traffic is served by the public endpoints within that pool.
| Chain | Status | Uptime 24h | 7 days | 30 days | p50 latency | p95 latency | Last checked |
|---|---|---|---|---|---|---|---|
Arbitrum One arbitrum_one |
Operational | 100% | 100% | 100% | 102 ms | 349 ms | less than a minute ago |
Base Mainnet base_mainnet |
Operational | 100% | 100% | 100% | 142 ms | 490 ms | less than a minute ago |
Ethereum Mainnet eth_mainnet |
Operational | 100% | 100% | 99.93% | 103 ms | 789 ms | less than a minute ago |
OP Mainnet optimism_mainnet |
Operational | 100% | 100% | 100% | 72 ms | 325 ms | less than a minute ago |
Polygon PoS polygon_pos |
Operational | 100% | 100% | 100% | 65 ms | 225 ms | less than a minute ago |
Solana Mainnet solana_mainnet |
Operational | 100% | 100% | 100% | 78 ms | 522 ms | less than a minute ago |
Last 30 days, day by day
Arbitrum One
Base Mainnet base_mainnet
Ethereum Mainnet eth_mainnet
OP Mainnet optimism_mainnet
Polygon PoS polygon_pos
Solana Mainnet solana_mainnet
Latency percentiles cover successful probes from the last 24 hours. AI agents can read
the same numbers via the flexyield://reliability MCP resource — see
/llms.txt.