perf(ecstore): reduce inline PUT commit overhead (#6033)

* perf(metrics): attribute PUT stage costs Co-Authored-By: heihutu <heihutu@gmail.com> * perf(ecstore): move PUT metadata during shuffle Co-Authored-By: heihutu <heihutu@gmail.com> * perf(s3): reuse PUT object lock state Co-Authored-By: heihutu <heihutu@gmail.com> * perf(ecstore): trim PUT metadata fanout clones Build per-disk PUT metadata only for committed writer slots, move the response metadata out of the fanout vector, and preserve fresh FileInfo shuffle semantics. Co-Authored-By: heihutu <heihutu@gmail.com> * perf(metrics): make PUT stage attribution opt-in Co-Authored-By: heihutu <heihutu@gmail.com> * perf(ecstore): commit inline PUT shards directly Co-Authored-By: heihutu <heihutu@gmail.com> * perf(ecstore): streamline rename staging cleanup Use the directory-specific removal operation for rename_data staging parents. This avoids a guaranteed failed file-removal probe on Unix-like hosts and lets Windows remove the empty directory directly while preserving best-effort non-empty handling. Co-Authored-By: heihutu <heihutu@gmail.com> * test(ecstore): cover inline PUT rename failures Cache the detailed stage metrics gate once per PUT and exercise exact-quorum and quorum-minus-one failures after inline shard encoding. Co-Authored-By: heihutu <heihutu@gmail.com> --------- Co-authored-by: heihutu <heihutu@gmail.com>
2026-08-20 11:32:19 +00:00 · 2026-08-13 10:04:20 +08:00
parent 846517625b
commit 3a0dbccc2e
9 changed files with 785 additions and 108 deletions
@@ -58,7 +58,7 @@ use std::sync::{
 /// When `false`, `record_put_object_path` and `record_put_object_stage_duration`
 /// become no-ops, and callers can skip the `Instant::now()` syscalls entirely.
 ///
-/// Set to `true` during startup when OTEL metric export is enabled.
+/// Enabled only through an explicit runtime opt-in.
 static PUT_STAGE_METRICS_ENABLED: AtomicBool = AtomicBool::new(false);
 static GET_STAGE_METRICS_ENABLED: AtomicBool = AtomicBool::new(false);

@@ -78,7 +78,7 @@ static METRICS_ENABLED: AtomicBool = AtomicBool::new(false);

 /// Enable or disable detailed per-stage PUT metrics.
 ///
-/// Called once during startup, typically gated by `rustfs_obs::observability_metric_enabled()`.
+/// Called once during startup after applying the detailed PUT attribution opt-in.
 pub fn set_put_stage_metrics_enabled(enabled: bool) {
    PUT_STAGE_METRICS_ENABLED.store(enabled, Ordering::Relaxed);
 }
@@ -103,6 +103,12 @@ pub fn put_stage_metrics_enabled() -> bool {
    PUT_STAGE_METRICS_ENABLED.load(Ordering::Relaxed)
 }

+/// Start a PUT-stage timer only when detailed PUT attribution is enabled.
+#[inline(always)]
+pub fn put_stage_timer() -> Option<std::time::Instant> {
+    put_stage_metrics_enabled().then(std::time::Instant::now)
+}
+
 #[inline(always)]
 pub fn get_stage_metrics_enabled() -> bool {
    GET_STAGE_METRICS_ENABLED.load(Ordering::Relaxed)
@@ -434,7 +440,7 @@ pub fn record_get_object_request_result(status: &str, duration_secs: f64) {
 /// Record PutObject request start.
 #[inline(always)]
 pub fn record_put_object_request_start(concurrent_requests: usize) {
-    if !put_stage_metrics_enabled() {
+    if !metrics_enabled() {
        return;
    }
    counter!("rustfs_io_put_object_requests_total").increment(1);
@@ -444,7 +450,7 @@ pub fn record_put_object_request_start(concurrent_requests: usize) {
 /// Record PutObject request result.
 #[inline(always)]
 pub fn record_put_object_request_result(status: &str, duration_secs: f64) {
-    if !put_stage_metrics_enabled() {
+    if !metrics_enabled() {
        return;
    }
    counter!("rustfs_io_put_object_request_results_total", "status" => status.to_string()).increment(1);
@@ -1905,7 +1911,7 @@ pub fn record_get_object(duration_ms: f64, size_bytes: i64) {
 /// * `zero_copy_eligible` - Whether the request was eligible for a zero-copy path
 #[inline(always)]
 pub fn record_put_object(duration_ms: f64, size_bytes: i64, zero_copy_eligible: bool) {
-    if !put_stage_metrics_enabled() {
+    if !metrics_enabled() {
        return;
    }
    counter!("rustfs_s3_put_object_total").increment(1);
@@ -2004,6 +2010,13 @@ pub fn record_put_object_stage_duration(stage: &'static str, duration_ms: f64) {
    histogram!("rustfs_s3_put_object_stage_duration_ms", "stage" => stage).record(duration_ms);
 }

+#[inline(always)]
+pub fn record_put_object_stage_duration_from(stage: &'static str, started_at: Option<std::time::Instant>) {
+    if let Some(started_at) = started_at {
+        record_put_object_stage_duration(stage, started_at.elapsed().as_secs_f64() * 1000.0);
+    }
+}
+
 /// Record generic internal operation stage duration (non-PUT paths).
 /// Use this for metacache walks, listing, lifecycle, and other background
 /// operations that are NOT part of the PUT object hot path.
@@ -2819,6 +2832,56 @@ mod tests {
        assert!(!put_stage_metrics_enabled());
    }

+    #[test]
+    fn put_stage_gate_does_not_disable_basic_put_metrics() {
+        let _guard = METRICS_FLAG_LOCK.lock().unwrap_or_else(|e| e.into_inner());
+        let recorder = DebuggingRecorder::new();
+        let snapshotter = recorder.snapshotter();
+
+        metrics::with_local_recorder(&recorder, || {
+            set_metrics_enabled(true);
+            set_put_stage_metrics_enabled(false);
+            record_put_object_request_start(1);
+            record_put_object_request_result("ok", 0.001);
+            record_put_object(1.0, 1024, false);
+            record_put_object_stage_duration("disabled_stage", 0.5);
+
+            set_put_stage_metrics_enabled(true);
+            record_put_object_stage_duration("enabled_stage", 0.5);
+
+            set_put_stage_metrics_enabled(false);
+            set_metrics_enabled(false);
+        });
+
+        let metrics = snapshotter.snapshot().into_vec();
+        assert!(metrics.iter().any(|(composite, _, _, _)| {
+            composite.kind() == MetricKind::Counter && composite.key().name() == "rustfs_s3_put_object_total"
+        }));
+        assert!(metrics.iter().any(|(composite, _, _, _)| {
+            composite.kind() == MetricKind::Counter && composite.key().name() == "rustfs_io_put_object_requests_total"
+        }));
+
+        let stages = metrics
+            .iter()
+            .filter(|(composite, _, _, _)| {
+                composite.kind() == MetricKind::Histogram && composite.key().name() == "rustfs_s3_put_object_stage_duration_ms"
+            })
+            .flat_map(|(composite, _, _, _)| composite.key().labels().map(|label| label.value().to_string()))
+            .collect::<Vec<_>>();
+        assert_eq!(stages, ["enabled_stage"]);
+    }
+
+    #[test]
+    fn test_put_stage_timer_follows_metrics_switch() {
+        let _guard = METRICS_FLAG_LOCK.lock().unwrap_or_else(|e| e.into_inner());
+        set_put_stage_metrics_enabled(false);
+        assert!(put_stage_timer().is_none());
+
+        set_put_stage_metrics_enabled(true);
+        assert!(put_stage_timer().is_some());
+        set_put_stage_metrics_enabled(false);
+    }
+
    #[test]
    fn test_record_get_object_path_and_stage() {
        let _guard = METRICS_FLAG_LOCK.lock().unwrap_or_else(|e| e.into_inner());