test(collector): stabilize TestMemStatsEquivalence against HeapReleased drift

The test read runtime.ReadMemStats and runtime/metrics separately and compared them within 5%, with the code itself noting it hoped no GC ran in between. HeapReleased is driven by the background scavenger and drifted between the two reads, failing intermittently (seen on Linux arm64). Force a full scavenge with debug.FreeOSMemory and disable GC for the measurement window so both reads observe the same runtime state.

Signed-off-by: Kemal Akkoyun <kemal.akkoyun@datadoghq.com>

Author: kakkoyun

prometheus/go_collector_latest_test.go

@@ -21,6 +21,7 @@ import (
 	"reflect"
 	"regexp"
 	"runtime"
+	"runtime/debug"
 	"runtime/metrics"
 	"strings"
 	"sync"
@@ -270,12 +271,16 @@ func TestMemStatsEquivalence(t *testing.T) {
 		samplesMap[descs[i].Name] = &samples[i]
 	}
 
-	// Force a GC cycle to try to reach a clean slate.
-	runtime.GC()
+	// Reach a stable slate and hold it for a single measurement window.
+	// FreeOSMemory runs a GC and returns freed memory to the OS, leaving the
+	// background scavenger nothing to release; disabling GC for the window
+	// keeps msReal and msFake observing the same runtime state. Without this,
+	// HeapReleased could drift between the two reads and fail the comparison.
+	debug.FreeOSMemory()
+	defer debug.SetGCPercent(debug.SetGCPercent(-1))
 
-	// Populate msReal.
+	// Populate msReal, then msFake back-to-back within the frozen window.
 	runtime.ReadMemStats(&msReal)
-	// Populate msFake and hope that no GC happened in between (:
 	metrics.Read(samples)
 
 	memStatsFromRM(&msFake, samplesMap)