[KEIAHNY]

PR Type

Enhancement

---

Description

• Extract shared get_up_windows window-detection logic

• Add get_window_cpu_average_vector CPU averaging helper

• Reuse unified logic in recycler and manager computations

• Remove resolved duplication issue from ISSUES.md

---

Diagram Walkthrough

flowchart LR
  A["get_recycler_count"] -- "calls" --> C["get_up_windows"]
  B["get_manager_count"] -- "calls" --> C
  C -- "uses" --> D["get_window_cpu_average_vector"]

Loading

File Walkthrough

	Relevant files
Enhancement	flow_threads.rs Unify recycler and manager window computation                        src/flow_threads.rs Extract duplicated window-scanning loop into generic get_up_windows with threshold and above parameters Add get_window_cpu_average_vector helper for per-window CPU usage averaging Refactor get_recycler_count and manager computation to consume the shared function Reset up flag each iteration and use WINDOWS instead of hardcoded 3 in tail check +66/-103
Documentation	ISSUES.md Remove resolved code duplication issue                                      ISSUES.md Remove resolved issue about duplicated recycler/manager mechanism +0/-1

---

[github-actions]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 3 🔵🔵🔵⚪⚪

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Dropped CPU check The old manager computation included a second gate: if avg > CPU_USAGE_MAX { up &= true; } else { up &= false; break; }. The unified get_up_windows only checks the slope threshold, so the manager path no longer requires CPU usage above CPU_USAGE_MAX for a window to count as "up". This silently changes get_manager_count behavior: windows that were previously rejected due to low manager CPU usage are now accepted, which can produce a lower (and incorrect) mgr_power minimum. In the new code the computed avg is also now unused, which corroborates that the check was lost rather than intentionally relocated. for current_window in &current_windows { let avg: f32 = avg_cpu_usage[(sm_counter * (avg_cpu_usage.len()/WINDOWS as usize)) .. ((sm_counter + 1) * (avg_cpu_usage.len()/WINDOWS as usize))] .iter().sum::<f32>() / (avg_cpu_usage.len()/WINDOWS as usize) as f32; sm_counter+=1; if (above && current_window > &threshold) || (!above && current_window < &threshold) { tmp_slope_vector.push(*current_window); up &= true; } else { up &= false;

[github-actions]

PR Code Suggestions ✨

Explore these optional code suggestions:

Category	Suggestion	Impact
Possible issue	Avoid out-of-bounds panic when slicing windows This slice panics if result.len() is not an exact multiple of WINDOWS, or if the float-equality-based break condition never matches before counter walks past the end of result. Use a bounds-checked slice and break cleanly when the data is exhausted, turning a potential panic into graceful termination. src/flow_threads.rs [86] - current_windows = result[(counter*WINDOWS) as usize..((counter+1)*WINDOWS) as usize].to_vec(); + let start = (counter * WINDOWS) as usize; + if start >= result.len() { + break; + } + let end = (((counter + 1) * WINDOWS) as usize).min(result.len()); + current_windows = result[start..end].to_vec(); Suggestion importance[1-10]: 7 __ Why: This addresses a real panic risk: the slice result[(counter*WINDOWS)..((counter+1)*WINDOWS)] will panic if result.len() is not an exact multiple of WINDOWS before the break condition triggers. The bounds-checked slice with a clean break converts a hard panic into graceful termination, matching the PR's refactored get_up_windows logic. Slightly reduced because the partial-window case could desync with the full-window avg_cpu_usage indexing, but the core fix is valid and important.	Medium
	Prevent NaN from empty slope vectors If current_windows is empty (e.g., when result is exhausted but the break condition hasn't triggered), the inner for loop never runs, up remains true, and an empty tmp_slope_vector is pushed. Downstream, dividing by tmp_slope_vector.len() as f64 yields NaN, and the later partial_cmp(...).expect(...) calls in get_recycler_count/get_manager_count will panic because NaN comparisons return None. Guard against empty windows before pushing. src/flow_threads.rs [113-115] - if up { + if up && !tmp_slope_vector.is_empty() { up_windows.push((counter, tmp_slope_vector.clone())); } Suggestion importance[1-10]: 3 __ Why: The downstream NaN/panic chain described (empty tmp_slope_vector producing NaN, then partial_cmp failing) is logically accurate, but an empty current_windows cannot actually occur here: result[(counter*WINDOWS)..((counter+1)*WINDOWS)] would panic on out-of-bounds rather than return an empty slice, so an empty tmp_slope_vector with up == true is effectively unreachable. The guard is a harmless defensive improvement with low practical impact.	Low

• Author self-review: I have reviewed the PR code suggestions, and addressed the relevant ones.

[KEIAHNY]

Closing - applying changes.