Baselines
Machine class: implementer / CI runners (2026-07).
Numbers below are CI-gated unit/integration targets plus methodology for Docker hosts.
Latency / concurrency targets
Section titled “Latency / concurrency targets”| Check | Target | Test |
|---|---|---|
| In-memory route lookup (1,000 routes) | p99 < 1 ms | momo_spec::routes::tests::thousand_route_lookups_within_latency_target |
| Concurrent cold start (100 waiters) | exactly 1 Docker start | momo_node::runtime::tests::concurrent_cold_single_start |
| Node transport failure | HTTP 503 / momo_node_unavailable |
momo_gateway::pool::tests::node_crash_maps_to_documented_gateway_error |
| Store watch outage | gateway keeps cached routes; watch_ok=false |
momo_gateway::tests::store_watch_unhealthy_keeps_serving_cached_routes |
Methodology (Docker host)
Section titled “Methodology (Docker host)”Warm path: container already Running; measure gateway proxy overhead.
Cold path: force docker stop then request until first successful HTTP body byte.
Repeat 3 times; report median.
# After momo node + gateway are up with a sample app:for i in 1 2 3; do curl -s -o /dev/null -w "warm_ms=%{time_total}\n" -H 'Host: api.example.com' http://127.0.0.1:8080/done