feat(hermes): add latency histograms for SSE/WS send paths #2977
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
![devin-ai-integration[bot]](https://avatars.githubusercontent.com/in/811515?s=60&v=4){kind=small}
Co-Authored-By: Tejas Badadare <[email protected]>
🤖 Devin AI EngineerI'll be helping with this pull request! Here's what you should know: ✅ I will automatically:
Note: I can only respond to comments from users who have write access to this repository. ⚙️ Control Options:
|
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Co-Authored-By: Tejas Badadare <[email protected]>
apps/hermes/server/src/api/ws.rs
Outdated
Comment on lines
512
to
530
| if let Some(received_at) = received_at_opt { | ||
| let pub_to_recv = (received_at - publish_time).max(0) as f64; | ||
| self.ws_state | ||
| .metrics | ||
| .publish_to_receive_latency | ||
| .observe(pub_to_recv); | ||
|
|
||
| let now_secs = std::time::SystemTime::now() | ||
| .duration_since(std::time::UNIX_EPOCH) | ||
| .ok() | ||
| .and_then(|d| i64::try_from(d.as_secs()).ok()) | ||
| .unwrap_or(received_at); | ||
| let recv_to_send = (now_secs - received_at).max(0) as f64; | ||
| self.ws_state | ||
| .metrics | ||
| .receive_to_ws_send_latency | ||
| .observe(recv_to_send); | ||
| } | ||
|
|
There was a problem hiding this comment.
This is in the wrong place, we should observe the latency after self.sender.flush()
…after flush Co-Authored-By: Tejas Badadare <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
Adds three new histogram metrics to the Hermes service to track latency in price feed updates:
publish_to_receive_latency: Time frompublish_timetoreceive_time(tracked for both SSE and WS)receive_to_ws_send_latency: Time fromreceive_timeto WebSocket send timereceive_to_sse_send_latency: Time fromreceive_timeto SSE send timeRationale
These metrics provide visibility into latency at different stages of the price feed pipeline, enabling monitoring and optimization of the Hermes service performance. The metrics help identify bottlenecks between price publication, reception, and delivery to clients.
Implementation Details
WsMetricsstruct to includepublish_to_receive_latencyandreceive_to_ws_send_latencyhistogramsSseMetricsstruct withreceive_to_sse_send_latencyhistogramApiStateto include SSE metricsArchitecture Notes
The
publish_to_receive_latencymetric is registered inWsMetricsbut observed by both endpoints. SSE accesses it viastate.ws.metrics.publish_to_receive_latencyto avoid duplicate registration.How has this been tested?
cargo build)Testing performed: Verified compilation success with
cargo build. No runtime testing of metrics collection was performed.Review Checklist
Please pay special attention to:
Metrics architecture: Is the cross-endpoint access pattern (
SSE -> WS metrics) acceptable or shouldpublish_to_receive_latencybe registered differently?Timing accuracy: Are the wall-clock measurements using
SystemTime::now()appropriate for latency tracking in hot paths?Performance impact: Review the overhead of adding histogram observations to every price update in both WS and SSE paths
Borrow checker solutions: Verify the approach of capturing
received_atandpublish_timebefore movingupdatestructs is soundHistogram buckets: Validate that the 0.1-20.0 second range with the chosen bucket distribution makes sense for expected latencies
Link to Devin run: https://app.devin.ai/sessions/e9af526709c3456c8bd160207b8ec67a
Requested by: Tejas Badadare ([email protected])