You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
M0: chat + media benchmark suite and the tokio relay baseline #2876
Milestone 0 of #2875. The whole thread-per-core project is gated on this: measure where relay CPU actually goes under a chat-shaped load before building anything.
Work items
Chat presets for moq-bench: 1:1 (chat.toml) and the 1:N pair (chat-pub.toml / chat-sub.toml).
Machine-readable output on the load side (--output, JSONL of cumulative counters).
moq-bench-host: production-safe host-side sampler (reads /proc only) for CPU, RSS, and context switches of the relay process, with a per-thread breakdown for later pinning verification.
Methodology docs: how to join the two JSONL streams into CPU per connection, CPU per message, and context switches per message.
Pick and document the reference hardware (dedicated box or fixed metal instance; cloud CI runners are too noisy).
Run the matrix against the current tokio relay: chat 1:1, chat 1:N (small and large fanout), media presets (hd, sd, audio), and record the baseline numbers.
Profile the dominant CPU sinks under the chat profile (perf flame graph on the reference box: syscall share, crypto, wakeups, memcpy).
Write down the target multiplier for the project based on the baseline.
Exit criteria
Baseline numbers for the current multi-thread tokio relay are written down, the dominant CPU sinks are identified, and the target multiplier is committed in #2875.
#2172 - moq-relay: expose a Prometheus /metrics endpoint for node traffic [merged] #2378 - feat(stats): record cumulative connection/broadcast/viewer time [open] #2430 - feat(stats): count datagrams in the model layer [merged] #2462 - build(rust): pin the toolchain and correct the MSRV claims [merged]
📝 Issue Planner
Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.
Create Plan
🧪 Issue enrichment is currently in open beta.
You can configure auto-planning by selecting labels in the issue_enrichment configuration.
To disable automatic issue enrichment, add the following to your .coderabbit.yaml:
issue_enrichment:
auto_enrich:
enabled: false
💬 Have feedback or questions? Drop into our discord!
Activity
coderabbitai commentedon Aug 15, 2026
🔗 Related PRs
#2172 - moq-relay: expose a Prometheus /metrics endpoint for node traffic [merged]
#2378 - feat(stats): record cumulative connection/broadcast/viewer time [open]
#2430 - feat(stats): count datagrams in the model layer [merged]
#2462 - build(rust): pin the toolchain and correct the MSRV claims [merged]
📝 Issue Planner
Check the box below or use the
@coderabbitai plancommand to generate an implementation plan and prompts that you can use with your favorite coding assistant.🧪 Issue enrichment is currently in open beta.
You can configure auto-planning by selecting labels in the issue_enrichment configuration.
To disable automatic issue enrichment, add the following to your
.coderabbit.yaml:💬 Have feedback or questions? Drop into our discord!