Metis Benchmarks

Benchmark Date: 08/24/2023

Providers included

  • Nodies

  • Blast API

  • Grove

Quick note - at the time of testing, we were not able to find many available Metis endpoints. We plan on revisiting testing in the future when a broader range of service providers becomes accessible.

RPC Method Tested

eth_getBlockByNumber

Flood Command

flood eth_getBlockByNumber <endpoint1> <endpoint2> <endpoint3> --rates 10 100 1000 10000 --duration 30

Benchmark Summary

Upon benchmarking the two Metis endpoints – BlastAPI, and Nodies – under varying request loads, we derived the following insights:

  • P50 Latency: Nodies consistently showcased the best performance across all request rates, with its latency not surpassing 6ms even at 10,000 requests per second (rps). BlastAPI had comparable latencies when the rate was 1,000 rps or lower. However, at 10,000 rps, BlastAPI experienced a significant spike, registering a latency of 119ms.

  • P90 Latency: Again, Nodies maintained the lead, keeping its latency under 9ms across all request loads. BlastAPI exhibited a gradual increase in latency, hitting 177ms at the highest load.

  • P99 Latency: At this percentile, the disparities became more pronounced. Nodies continued its consistent performance, with latencies not going above 10ms. BlastAPI's performance curve remained relatively steady, reaching a latency of about 1,185ms at 10,000 rps. However,

In conclusion, the Nodies endpoint demonstrated superior latency performance across all percentiles and request loads. BlastAPI exhibited comparable latencies at lower loads, but its performance diverged significantly at higher loads.

Latency Benchmarks

Metrics: Latency percentiles (P50, P90, P99) against load rates (10, 100, 1000, and 10000 rps).

P50 Latency (in ms, lower is better):

P90 Latency (in ms, lower is better):

P99 Latency (in ms. lower is better):

Last updated