How to publish a useful RPC / gRPC benchmark

[admin]

A result without a method is hard to use. Please include:

• Test date, region, host specifications and network conditions.
• Provider, plan, protocol, methods and commitment level. Redact API credentials.
• Tool version or commit, configuration, sample size, concurrency and warm-up.
• p50/p95/p99 latency, errors, timeouts and rate limits. Distinguish request latency from event-delivery latency.
• Raw anonymized results and commands that another person can run.
• Limitations, provider relationships and paid/sponsored access.

Do not present a single run as a universal ranking. Compare equivalent workloads and respect provider limits. This is a reporting template, not a published benchmark result.