pcool: Client-side end-to-end latency measurement
- Dominant language
- Rust
- Stars
- 110
- Forks
- 71
- Avg merge
- 3d 4h
- Merged PRs (30d)
- 129
Description
> **Epic** — break into sub-tasks when work begins.
Server-side latency measurements (#10750) are not observable by clients. Implement client-side latency tracking in the SDK so clients can independently select the best fullnode.
Depends on: #10750
---
### Suggested sub-tasks (create when starting)
- Design per-fullnode score structure and decay algorithm (mirror #10750 model)
- Implement submission-to-effects latency recording in Rust SDK
- Implement weighted fullnode selection based on recorded scores
- TypeScript SDK equivalent (if coverage needed)
- Tests and benchmarks
Contributor guide
Research direction
This is an epic with no files or tests named; start by reading issue #10750 and identifying the Rust SDK entry points for submission-to-effects latency and fullnode selection. Break the work into the listed score, recording, selection, TypeScript, and test/benchmark sub-tasks, with completion defined by client-side measurements and weighted selection covered by tests.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust, typescript
- Domain
- api
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 25/100