Response Time Monitor
Upstream latency percentiles for Solana RPC methods
Upstream latency is the time spent by SVS infrastructure processing each RPC request, excluding network transit between you and our edge. It's the best signal for evaluating service performance independent of client network conditions.
These percentiles are aggregated from every RPC request our clients make against our infrastructure. This is live production traffic, not synthetic benchmarks. Latency is pooled across all upstream hosts in the selected region.
p50 is the median request; p95 and p99 show the slower tail. Lower is better.
Live RPC tx-landing metrics in Discord
Real-time transaction-landing telemetry posted to our Discord. Useful in tandem with the upstream latency above. Open the channel to subscribe.
Transaction landing snapshot
Latest read from our metrics bot. Lightspeed sampled hourly; SWQoS / Unstaked sampled every 5 minutes.
Lightspeed
- Current
- 0.205 s
- 24h average
- 0.257 s
- 24h P95
- 0.533 s
- 24h P99
- 0.606 s
- 24h success
- 22 / 22 (100%)
- 7d success
- 151 / 151 (100%)
SWQoS (staked)
- Current
- 0.359 s
- 24h average
- 0.275 s
- 24h P95
- 0.694 s
- 24h P99
- 1.005 s
- 24h success
- 259 / 259 (100%)
- 7d success
- 1,808 / 1,808 (100%)
Unstaked
- Current
- 1.533 s
- 24h average
- 0.672 s
- 24h P95
- 2.455 s
- 24h P99
- 3.353 s
- 24h success
- 256 / 259 (98.8%)
- 7d success
- 1,800 / 1,808 (99.6%)
Live updates resume when the metrics bot is back online. Subscribe to the Discord channel to follow.
Latency — p50 / p95 / p99
Times shown in your local timezone. Source resolution is ~1ms; values below that are shown as <1ms.
All Methods — Atlanta (ATL)
Latest 5-minute bucket per method. Click a row to chart its history.
No data available