fix: use 5-minute load average for resource alerts to reduce false positives

The 1-minute load average spikes during normal build/test runs,
triggering false "critically low resources" alerts. The 5-minute
average smooths transient spikes while still catching sustained
resource pressure.

Spec: specs/load-alert-5min-average.md
Signed-off-by: Jose Alekhinne <jose@parlakisik.com>

Author: josealekhine

internal/cli/doctor/doctor_test.go

@@ -403,7 +403,7 @@ func TestAddResourceResults_DangerMapsToError(t *testing.T) {
 			Supported:  true,
 		},
 		Load: sysinfo.LoadInfo{
-			Load1:     12.0,
+			Load5:     12.0,
 			NumCPU:    8, // 1.5x -> DANGER
 			Supported: true,
 		},
@@ -479,7 +479,7 @@ func TestAddResourceResults_MessageFormat(t *testing.T) {
 		},
 		Disk: sysinfo.DiskInfo{Supported: false},
 		Load: sysinfo.LoadInfo{
-			Load1:     2.0,
+			Load5:     2.0,
 			NumCPU:    8,
 			Supported: true,
 		},

internal/sysinfo/threshold.go

@@ -96,9 +96,9 @@ func Evaluate(snap Snapshot) []ResourceAlert {
 		}
 	}
 
-	// Load (1m)
+	// Load (5m) — 5-minute average smooths transient build/test spikes.
 	if snap.Load.Supported && snap.Load.NumCPU > 0 {
-		ratio := snap.Load.Load1 / float64(snap.Load.NumCPU)
+		ratio := snap.Load.Load5 / float64(snap.Load.NumCPU)
 		msg := fmt.Sprintf(desc.Text(text.DescKeyResourcesAlertLoad), ratio)
 		if ratio >= stats.ThresholdLoadDangerRatio {
 			alerts = append(alerts, ResourceAlert{

internal/sysinfo/threshold_test.go

@@ -122,7 +122,7 @@ func TestEvaluate_DiskBoundaries(t *testing.T) {
 func TestEvaluate_LoadBoundaries(t *testing.T) {
 	tests := []struct {
 		name    string
-		load1   float64
+		load5   float64
 		ncpu    int
 		wantSev Severity
 		wantN   int
@@ -137,7 +137,7 @@ func TestEvaluate_LoadBoundaries(t *testing.T) {
 		t.Run(tt.name, func(t *testing.T) {
 			snap := Snapshot{
 				Load: LoadInfo{
-					Load1:     tt.load1,
+					Load5:     tt.load5,
 					NumCPU:    tt.ncpu,
 					Supported: true,
 				},
@@ -171,7 +171,7 @@ func TestEvaluate_AllDanger(t *testing.T) {
 			Supported:  true,
 		},
 		Load: LoadInfo{
-			Load1:     12.0,
+			Load5:     12.0,
 			NumCPU:    8,
 			Supported: true,
 		},

specs/load-alert-5min-average.md

@@ -0,0 +1,19 @@
+# Load Alert: Use 5-Minute Average
+
+## Problem
+
+The resource load alert uses the 1-minute load average, which spikes
+during normal build/test runs and produces false positives. A single
+`make test` run triggers "critically low resources" even when the
+system is healthy.
+
+## Approach
+
+Switch from `Load1` to `Load5` in `sysinfo.Evaluate`. The 5-minute
+average smooths transient build/test spikes while still catching
+sustained resource pressure.
+
+## Non-Goals
+
+- Changing the threshold ratios (0.8 warn, 1.5 danger)
+- Adding configurable averaging windows