refactor(obs): hide replication stats handle (#4149)

crates/obs/src/metrics/mod.rs

@@ -31,9 +31,9 @@ pub use scheduler::{
     init_metrics_runtime, metrics_runtime_controller_snapshot, metrics_runtime_status_snapshot,
 };
 pub(crate) use storage_api::metrics::{
-    BucketOperations, BucketOptions, ObsBucketBandwidthMonitor, ObsEcstoreResult, ObsReplicationStats, ObsStore, StorageAdminApi,
-    obs_expiry_state_handle, obs_get_global_bucket_monitor, obs_get_global_replication_stats, obs_get_quota_config,
+    BucketOperations, BucketOptions, ObsBucketBandwidthMonitor, ObsEcstoreResult, ObsStore, StorageAdminApi,
+    obs_bucket_replication_stats_snapshot, obs_expiry_state_handle, obs_get_global_bucket_monitor, obs_get_quota_config,
     obs_get_total_usable_capacity, obs_get_total_usable_capacity_free, obs_is_disk_compression_enabled,
-    obs_load_compression_total_from_memory, obs_load_data_usage_from_backend, obs_resolve_object_store_handle,
-    obs_transition_state_handle,
+    obs_load_compression_total_from_memory, obs_load_data_usage_from_backend, obs_replication_site_stats_snapshot,
+    obs_resolve_object_store_handle, obs_transition_state_handle,
 };

crates/obs/src/metrics/runtime_sources.rs

@@ -13,14 +13,11 @@
 // limitations under the License.
 
 use crate::metrics::{
-    ObsBucketBandwidthMonitor, ObsReplicationStats, obs_expiry_state_handle, obs_get_global_bucket_monitor,
-    obs_get_global_replication_stats, obs_transition_state_handle,
+    ObsBucketBandwidthMonitor, obs_expiry_state_handle, obs_get_global_bucket_monitor, obs_transition_state_handle,
 };
 use rustfs_iam::{get_global_iam_sys, oidc::oidc_plugin_authn_metrics_snapshot};
 use std::sync::Arc;
 
-pub(crate) type ObsReplicationStatsHandle = Arc<ObsReplicationStats>;
-
 #[derive(Debug, Clone, Copy, PartialEq, Eq)]
 pub(crate) struct ObsIlmRuntimeSnapshot {
     pub(crate) expiry_pending_tasks: u64,
@@ -63,10 +60,6 @@ pub(crate) fn bucket_monitor_available() -> bool {
     bucket_monitor_handle().is_some()
 }
 
-pub(crate) fn replication_stats_handle() -> Option<ObsReplicationStatsHandle> {
-    obs_get_global_replication_stats()
-}
-
 pub(crate) async fn ilm_runtime_snapshot() -> ObsIlmRuntimeSnapshot {
     let expiry_state = obs_expiry_state_handle();
     let transition_state = obs_transition_state_handle();

crates/obs/src/metrics/stats_collector.rs

@@ -26,20 +26,19 @@ use crate::metrics::collectors::{
     DriveCountStats, DriveDetailedStats, ErasureSetStats, HostNetworkStats, IamStats, IlmStats, MemoryStats, NetworkStats,
     ProcessStats, ProcessStatusType, ReplicationStats, ResourceStats, ScannerStats,
 };
-use crate::metrics::runtime_sources::{
-    ObsIlmRuntimeSnapshot, bucket_monitor_handle, iam_metrics_snapshot, ilm_runtime_snapshot, replication_stats_handle,
-};
+use crate::metrics::runtime_sources::{ObsIlmRuntimeSnapshot, bucket_monitor_handle, iam_metrics_snapshot, ilm_runtime_snapshot};
 use crate::metrics::{
-    BucketOperations, BucketOptions, ObsEcstoreResult, ObsStore, StorageAdminApi, obs_get_quota_config,
-    obs_get_total_usable_capacity, obs_get_total_usable_capacity_free, obs_load_compression_total_from_memory,
-    obs_load_data_usage_from_backend, obs_resolve_object_store_handle,
+    BucketOperations, BucketOptions, ObsEcstoreResult, ObsStore, StorageAdminApi, obs_bucket_replication_stats_snapshot,
+    obs_get_quota_config, obs_get_total_usable_capacity, obs_get_total_usable_capacity_free,
+    obs_load_compression_total_from_memory, obs_load_data_usage_from_backend, obs_replication_site_stats_snapshot,
+    obs_resolve_object_store_handle,
 };
 use chrono::Utc;
 use rustfs_common::heal_channel::HealScanMode;
 use rustfs_common::metrics::global_metrics;
 use rustfs_io_metrics::internode_metrics::global_internode_metrics;
 use rustfs_io_metrics::{ProcessStatusSnapshot, snapshot_process_resource_and_system};
-use std::{collections::HashMap, sync::Arc, time::Duration};
+use std::{collections::HashMap, sync::Arc};
 use sysinfo::{Networks, System};
 use tracing::{instrument, warn};
 
@@ -79,14 +78,6 @@ fn usize_to_u64_saturating(value: usize) -> u64 {
     u64::try_from(value).unwrap_or(u64::MAX)
 }
 
-fn i64_to_u64_floor_zero(value: i64) -> u64 {
-    u64::try_from(value.max(0)).unwrap_or(0)
-}
-
-fn i32_to_u64_floor_zero(value: i32) -> u64 {
-    u64::try_from(value.max(0)).unwrap_or(0)
-}
-
 async fn load_obs_data_usage_from_backend(store: Arc<ObsStore>) -> ObsEcstoreResult<ObsDataUsageInfo> {
     let data_usage = obs_load_data_usage_from_backend(store).await?;
 
@@ -158,116 +149,67 @@ async fn obs_ilm_runtime_snapshot() -> ObsIlmRuntimeSnapshot {
 }
 
 async fn obs_bucket_replication_detail_stats() -> Vec<BucketReplicationStats> {
-    let Some(stats) = replication_stats_handle() else {
-        return Vec::new();
-    };
-
-    let all_bucket_stats = stats.get_all().await;
-    let mut buckets = Vec::with_capacity(all_bucket_stats.len());
-
-    for (bucket, bucket_stats) in all_bucket_stats {
-        let proxy = stats.get_proxy_stats(&bucket).await;
-        let mut total_failed_bytes = 0u64;
-        let mut total_failed_count = 0u64;
-        let mut last_min_failed_bytes = 0u64;
-        let mut last_min_failed_count = 0u64;
-        let mut last_hour_failed_bytes = 0u64;
-        let mut last_hour_failed_count = 0u64;
-        let mut sent_bytes = 0u64;
-        let mut sent_count = 0u64;
-        let mut targets = Vec::with_capacity(bucket_stats.stats.len());
-
-        for (target_arn, target_stats) in bucket_stats.stats {
-            total_failed_bytes += i64_to_u64_floor_zero(target_stats.fail_stats.size);
-            total_failed_count += i64_to_u64_floor_zero(target_stats.fail_stats.count);
-
-            let last_min = target_stats.fail_stats.recent_since(Duration::from_secs(60));
-            last_min_failed_bytes += i64_to_u64_floor_zero(last_min.size);
-            last_min_failed_count += i64_to_u64_floor_zero(last_min.count);
-
-            let last_hour = target_stats.fail_stats.recent_since(Duration::from_secs(60 * 60));
-            last_hour_failed_bytes += i64_to_u64_floor_zero(last_hour.size);
-            last_hour_failed_count += i64_to_u64_floor_zero(last_hour.count);
-
-            sent_bytes += i64_to_u64_floor_zero(target_stats.replicated_size);
-            sent_count += i64_to_u64_floor_zero(target_stats.replicated_count);
-
-            targets.push(BucketReplicationTargetStats {
-                target_arn,
-                bandwidth_limit_bytes_per_sec: i64_to_u64_floor_zero(target_stats.bandwidth_limit_bytes_per_sec),
-                current_bandwidth_bytes_per_sec: target_stats.current_bandwidth_bytes_per_sec,
-                latency_ms: target_stats.latency.curr,
-            });
-        }
-
-        buckets.push(BucketReplicationStats {
-            bucket,
-            total_failed_bytes,
-            total_failed_count,
-            last_min_failed_bytes,
-            last_min_failed_count,
-            last_hour_failed_bytes,
-            last_hour_failed_count,
-            sent_bytes,
-            sent_count,
-            proxied_get_requests_total: i64_to_u64_floor_zero(proxy.get_total),
-            proxied_get_requests_failures: i64_to_u64_floor_zero(proxy.get_failed),
-            proxied_head_requests_total: i64_to_u64_floor_zero(proxy.head_total),
-            proxied_head_requests_failures: i64_to_u64_floor_zero(proxy.head_failed),
-            proxied_put_requests_total: i64_to_u64_floor_zero(proxy.put_total),
-            proxied_put_requests_failures: i64_to_u64_floor_zero(proxy.put_failed),
-            proxied_put_tagging_requests_total: i64_to_u64_floor_zero(proxy.put_tag_total),
-            proxied_put_tagging_requests_failures: i64_to_u64_floor_zero(proxy.put_tag_failed),
-            proxied_get_tagging_requests_total: i64_to_u64_floor_zero(proxy.get_tag_total),
-            proxied_get_tagging_requests_failures: i64_to_u64_floor_zero(proxy.get_tag_failed),
-            proxied_delete_tagging_requests_total: i64_to_u64_floor_zero(proxy.delete_tag_total),
-            proxied_delete_tagging_requests_failures: i64_to_u64_floor_zero(proxy.delete_tag_failed),
-            targets,
-        });
-    }
-
-    buckets
+    obs_bucket_replication_stats_snapshot()
+        .await
+        .into_iter()
+        .map(|stats| BucketReplicationStats {
+            bucket: stats.bucket,
+            total_failed_bytes: stats.total_failed_bytes,
+            total_failed_count: stats.total_failed_count,
+            last_min_failed_bytes: stats.last_min_failed_bytes,
+            last_min_failed_count: stats.last_min_failed_count,
+            last_hour_failed_bytes: stats.last_hour_failed_bytes,
+            last_hour_failed_count: stats.last_hour_failed_count,
+            sent_bytes: stats.sent_bytes,
+            sent_count: stats.sent_count,
+            proxied_get_requests_total: stats.proxied_get_requests_total,
+            proxied_get_requests_failures: stats.proxied_get_requests_failures,
+            proxied_head_requests_total: stats.proxied_head_requests_total,
+            proxied_head_requests_failures: stats.proxied_head_requests_failures,
+            proxied_put_requests_total: stats.proxied_put_requests_total,
+            proxied_put_requests_failures: stats.proxied_put_requests_failures,
+            proxied_put_tagging_requests_total: stats.proxied_put_tagging_requests_total,
+            proxied_put_tagging_requests_failures: stats.proxied_put_tagging_requests_failures,
+            proxied_get_tagging_requests_total: stats.proxied_get_tagging_requests_total,
+            proxied_get_tagging_requests_failures: stats.proxied_get_tagging_requests_failures,
+            proxied_delete_tagging_requests_total: stats.proxied_delete_tagging_requests_total,
+            proxied_delete_tagging_requests_failures: stats.proxied_delete_tagging_requests_failures,
+            targets: stats
+                .targets
+                .into_iter()
+                .map(|target| BucketReplicationTargetStats {
+                    target_arn: target.target_arn,
+                    bandwidth_limit_bytes_per_sec: target.bandwidth_limit_bytes_per_sec,
+                    current_bandwidth_bytes_per_sec: target.current_bandwidth_bytes_per_sec,
+                    latency_ms: target.latency_ms,
+                })
+                .collect(),
+        })
+        .collect()
 }
 
 async fn obs_site_replication_stats() -> ReplicationStats {
-    let Some(stats) = replication_stats_handle() else {
-        return ReplicationStats::default();
-    };
-
-    let site_metrics = stats.get_sr_metrics_for_node().await;
     let current_data_transfer_rate = obs_bucket_replication_bandwidth_stats()
         .into_iter()
         .flatten()
         .map(|stat| stat.current_bandwidth_bytes_per_sec)
         .sum::<f64>();
-
-    let all_bucket_stats = stats.get_all().await;
-    let average_data_transfer_rate = all_bucket_stats
-        .values()
-        .flat_map(|bucket| bucket.stats.values())
-        .map(|stat| stat.xfer_rate_lrg.avg + stat.xfer_rate_sml.avg)
-        .sum::<f64>();
-    let max_data_transfer_rate = all_bucket_stats
-        .values()
-        .flat_map(|bucket| bucket.stats.values())
-        .map(|stat| stat.xfer_rate_lrg.peak + stat.xfer_rate_sml.peak)
-        .sum::<f64>();
-    let recent_backlog_count = stats.mrf_stats.values().copied().filter(|value| *value > 0).sum::<i64>();
+    let stats = obs_replication_site_stats_snapshot(current_data_transfer_rate).await;
 
     ReplicationStats {
-        average_active_workers: site_metrics.active_workers.avg,
-        average_queued_bytes: site_metrics.queued.avg.bytes,
-        average_queued_count: site_metrics.queued.avg.count,
-        average_data_transfer_rate,
-        active_workers: i32_to_u64_floor_zero(site_metrics.active_workers.curr),
-        current_data_transfer_rate,
-        last_minute_queued_bytes: i64_to_u64_floor_zero(site_metrics.queued.last_minute.bytes),
-        last_minute_queued_count: i64_to_u64_floor_zero(site_metrics.queued.last_minute.count),
-        max_active_workers: i32_to_u64_floor_zero(site_metrics.active_workers.max),
-        max_queued_bytes: i64_to_u64_floor_zero(site_metrics.queued.max.bytes),
-        max_queued_count: i64_to_u64_floor_zero(site_metrics.queued.max.count),
-        max_data_transfer_rate,
-        recent_backlog_count: i64_to_u64_floor_zero(recent_backlog_count),
+        average_active_workers: stats.average_active_workers,
+        average_queued_bytes: stats.average_queued_bytes,
+        average_queued_count: stats.average_queued_count,
+        average_data_transfer_rate: stats.average_data_transfer_rate,
+        active_workers: stats.active_workers,
+        current_data_transfer_rate: stats.current_data_transfer_rate,
+        last_minute_queued_bytes: stats.last_minute_queued_bytes,
+        last_minute_queued_count: stats.last_minute_queued_count,
+        max_active_workers: stats.max_active_workers,
+        max_queued_bytes: stats.max_queued_bytes,
+        max_queued_count: stats.max_queued_count,
+        max_data_transfer_rate: stats.max_data_transfer_rate,
+        recent_backlog_count: stats.recent_backlog_count,
     }
 }
 

crates/obs/src/metrics/storage_api.rs

@@ -12,11 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::time::Duration;
+
 pub(crate) use rustfs_ecstore::api::bucket::bandwidth::monitor::Monitor as ObsBucketBandwidthMonitor;
 pub(crate) use rustfs_ecstore::api::bucket::metadata_sys::get_quota_config as obs_get_quota_config;
-pub(crate) use rustfs_ecstore::api::bucket::replication::{
-    ReplicationStats as ObsReplicationStats, get_global_replication_stats as obs_get_global_replication_stats,
-};
+use rustfs_ecstore::api::bucket::replication::get_global_replication_stats;
 pub(crate) use rustfs_ecstore::api::capacity::{
     get_total_usable_capacity as obs_get_total_usable_capacity,
     get_total_usable_capacity_free as obs_get_total_usable_capacity_free,
@@ -32,13 +32,194 @@ pub(crate) use rustfs_ecstore::api::runtime::{
 pub(crate) use rustfs_ecstore::api::storage::ECStore as ObsStore;
 use rustfs_storage_api as storage_contracts;
 
+#[derive(Debug, Clone, PartialEq)]
+pub(crate) struct ObsBucketReplicationTargetStatsSnapshot {
+    pub(crate) target_arn: String,
+    pub(crate) bandwidth_limit_bytes_per_sec: u64,
+    pub(crate) current_bandwidth_bytes_per_sec: f64,
+    pub(crate) latency_ms: f64,
+}
+
+#[derive(Debug, Clone, PartialEq)]
+pub(crate) struct ObsBucketReplicationStatsSnapshot {
+    pub(crate) bucket: String,
+    pub(crate) total_failed_bytes: u64,
+    pub(crate) total_failed_count: u64,
+    pub(crate) last_min_failed_bytes: u64,
+    pub(crate) last_min_failed_count: u64,
+    pub(crate) last_hour_failed_bytes: u64,
+    pub(crate) last_hour_failed_count: u64,
+    pub(crate) sent_bytes: u64,
+    pub(crate) sent_count: u64,
+    pub(crate) proxied_get_requests_total: u64,
+    pub(crate) proxied_get_requests_failures: u64,
+    pub(crate) proxied_head_requests_total: u64,
+    pub(crate) proxied_head_requests_failures: u64,
+    pub(crate) proxied_put_requests_total: u64,
+    pub(crate) proxied_put_requests_failures: u64,
+    pub(crate) proxied_put_tagging_requests_total: u64,
+    pub(crate) proxied_put_tagging_requests_failures: u64,
+    pub(crate) proxied_get_tagging_requests_total: u64,
+    pub(crate) proxied_get_tagging_requests_failures: u64,
+    pub(crate) proxied_delete_tagging_requests_total: u64,
+    pub(crate) proxied_delete_tagging_requests_failures: u64,
+    pub(crate) targets: Vec<ObsBucketReplicationTargetStatsSnapshot>,
+}
+
+#[derive(Debug, Clone, Default, PartialEq)]
+pub(crate) struct ObsReplicationSiteStatsSnapshot {
+    pub(crate) average_active_workers: f64,
+    pub(crate) average_queued_bytes: i64,
+    pub(crate) average_queued_count: i64,
+    pub(crate) average_data_transfer_rate: f64,
+    pub(crate) active_workers: u64,
+    pub(crate) current_data_transfer_rate: f64,
+    pub(crate) last_minute_queued_bytes: u64,
+    pub(crate) last_minute_queued_count: u64,
+    pub(crate) max_active_workers: u64,
+    pub(crate) max_queued_bytes: u64,
+    pub(crate) max_queued_count: u64,
+    pub(crate) max_data_transfer_rate: f64,
+    pub(crate) recent_backlog_count: u64,
+}
+
+fn i64_to_u64_floor_zero(value: i64) -> u64 {
+    u64::try_from(value.max(0)).unwrap_or(0)
+}
+
+fn i32_to_u64_floor_zero(value: i32) -> u64 {
+    u64::try_from(value.max(0)).unwrap_or(0)
+}
+
+pub(crate) async fn obs_bucket_replication_stats_snapshot() -> Vec<ObsBucketReplicationStatsSnapshot> {
+    let Some(stats) = get_global_replication_stats() else {
+        return Vec::new();
+    };
+
+    let all_bucket_stats = stats.get_all().await;
+    let mut buckets = Vec::with_capacity(all_bucket_stats.len());
+
+    for (bucket, bucket_stats) in all_bucket_stats {
+        let proxy = stats.get_proxy_stats(&bucket).await;
+        let mut total_failed_bytes = 0u64;
+        let mut total_failed_count = 0u64;
+        let mut last_min_failed_bytes = 0u64;
+        let mut last_min_failed_count = 0u64;
+        let mut last_hour_failed_bytes = 0u64;
+        let mut last_hour_failed_count = 0u64;
+        let mut sent_bytes = 0u64;
+        let mut sent_count = 0u64;
+        let mut targets = Vec::with_capacity(bucket_stats.stats.len());
+
+        for (target_arn, target_stats) in bucket_stats.stats {
+            total_failed_bytes = total_failed_bytes.saturating_add(i64_to_u64_floor_zero(target_stats.fail_stats.size));
+            total_failed_count = total_failed_count.saturating_add(i64_to_u64_floor_zero(target_stats.fail_stats.count));
+
+            let last_min = target_stats.fail_stats.recent_since(Duration::from_secs(60));
+            last_min_failed_bytes = last_min_failed_bytes.saturating_add(i64_to_u64_floor_zero(last_min.size));
+            last_min_failed_count = last_min_failed_count.saturating_add(i64_to_u64_floor_zero(last_min.count));
+
+            let last_hour = target_stats.fail_stats.recent_since(Duration::from_secs(60 * 60));
+            last_hour_failed_bytes = last_hour_failed_bytes.saturating_add(i64_to_u64_floor_zero(last_hour.size));
+            last_hour_failed_count = last_hour_failed_count.saturating_add(i64_to_u64_floor_zero(last_hour.count));
+
+            sent_bytes = sent_bytes.saturating_add(i64_to_u64_floor_zero(target_stats.replicated_size));
+            sent_count = sent_count.saturating_add(i64_to_u64_floor_zero(target_stats.replicated_count));
+
+            targets.push(ObsBucketReplicationTargetStatsSnapshot {
+                target_arn,
+                bandwidth_limit_bytes_per_sec: i64_to_u64_floor_zero(target_stats.bandwidth_limit_bytes_per_sec),
+                current_bandwidth_bytes_per_sec: target_stats.current_bandwidth_bytes_per_sec,
+                latency_ms: target_stats.latency.curr,
+            });
+        }
+
+        buckets.push(ObsBucketReplicationStatsSnapshot {
+            bucket,
+            total_failed_bytes,
+            total_failed_count,
+            last_min_failed_bytes,
+            last_min_failed_count,
+            last_hour_failed_bytes,
+            last_hour_failed_count,
+            sent_bytes,
+            sent_count,
+            proxied_get_requests_total: i64_to_u64_floor_zero(proxy.get_total),
+            proxied_get_requests_failures: i64_to_u64_floor_zero(proxy.get_failed),
+            proxied_head_requests_total: i64_to_u64_floor_zero(proxy.head_total),
+            proxied_head_requests_failures: i64_to_u64_floor_zero(proxy.head_failed),
+            proxied_put_requests_total: i64_to_u64_floor_zero(proxy.put_total),
+            proxied_put_requests_failures: i64_to_u64_floor_zero(proxy.put_failed),
+            proxied_put_tagging_requests_total: i64_to_u64_floor_zero(proxy.put_tag_total),
+            proxied_put_tagging_requests_failures: i64_to_u64_floor_zero(proxy.put_tag_failed),
+            proxied_get_tagging_requests_total: i64_to_u64_floor_zero(proxy.get_tag_total),
+            proxied_get_tagging_requests_failures: i64_to_u64_floor_zero(proxy.get_tag_failed),
+            proxied_delete_tagging_requests_total: i64_to_u64_floor_zero(proxy.delete_tag_total),
+            proxied_delete_tagging_requests_failures: i64_to_u64_floor_zero(proxy.delete_tag_failed),
+            targets,
+        });
+    }
+
+    buckets
+}
+
+pub(crate) async fn obs_replication_site_stats_snapshot(current_data_transfer_rate: f64) -> ObsReplicationSiteStatsSnapshot {
+    let Some(stats) = get_global_replication_stats() else {
+        return ObsReplicationSiteStatsSnapshot::default();
+    };
+
+    let site_metrics = stats.get_sr_metrics_for_node().await;
+    let all_bucket_stats = stats.get_all().await;
+    // These fields keep the existing metric semantics: cluster-wide sums across bucket targets.
+    let average_data_transfer_rate = all_bucket_stats
+        .values()
+        .flat_map(|bucket| bucket.stats.values())
+        .map(|stat| stat.xfer_rate_lrg.avg + stat.xfer_rate_sml.avg)
+        .sum::<f64>();
+    let max_data_transfer_rate = all_bucket_stats
+        .values()
+        .flat_map(|bucket| bucket.stats.values())
+        .map(|stat| stat.xfer_rate_lrg.peak + stat.xfer_rate_sml.peak)
+        .sum::<f64>();
+    let recent_backlog_count = stats.mrf_stats.values().copied().filter(|value| *value > 0).sum::<i64>();
+
+    ObsReplicationSiteStatsSnapshot {
+        average_active_workers: site_metrics.active_workers.avg,
+        average_queued_bytes: site_metrics.queued.avg.bytes,
+        average_queued_count: site_metrics.queued.avg.count,
+        average_data_transfer_rate,
+        active_workers: i32_to_u64_floor_zero(site_metrics.active_workers.curr),
+        current_data_transfer_rate,
+        last_minute_queued_bytes: i64_to_u64_floor_zero(site_metrics.queued.last_minute.bytes),
+        last_minute_queued_count: i64_to_u64_floor_zero(site_metrics.queued.last_minute.count),
+        max_active_workers: i32_to_u64_floor_zero(site_metrics.active_workers.max),
+        max_queued_bytes: i64_to_u64_floor_zero(site_metrics.queued.max.bytes),
+        max_queued_count: i64_to_u64_floor_zero(site_metrics.queued.max.count),
+        max_data_transfer_rate,
+        recent_backlog_count: i64_to_u64_floor_zero(recent_backlog_count),
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn obs_replication_numeric_conversions_floor_negative_values() {
+        assert_eq!(i64_to_u64_floor_zero(-1), 0);
+        assert_eq!(i64_to_u64_floor_zero(42), 42);
+        assert_eq!(i32_to_u64_floor_zero(-1), 0);
+        assert_eq!(i32_to_u64_floor_zero(42), 42);
+    }
+}
+
 pub(crate) mod metrics {
     pub(crate) use super::storage_contracts::{BucketOperations, BucketOptions, StorageAdminApi};
 
     pub(crate) use super::{
-        ObsBucketBandwidthMonitor, ObsEcstoreResult, ObsReplicationStats, ObsStore, obs_expiry_state_handle,
-        obs_get_global_bucket_monitor, obs_get_global_replication_stats, obs_get_quota_config, obs_get_total_usable_capacity,
-        obs_get_total_usable_capacity_free, obs_is_disk_compression_enabled, obs_load_compression_total_from_memory,
-        obs_load_data_usage_from_backend, obs_resolve_object_store_handle, obs_transition_state_handle,
+        ObsBucketBandwidthMonitor, ObsEcstoreResult, ObsStore, obs_bucket_replication_stats_snapshot, obs_expiry_state_handle,
+        obs_get_global_bucket_monitor, obs_get_quota_config, obs_get_total_usable_capacity, obs_get_total_usable_capacity_free,
+        obs_is_disk_compression_enabled, obs_load_compression_total_from_memory, obs_load_data_usage_from_backend,
+        obs_replication_site_stats_snapshot, obs_resolve_object_store_handle, obs_transition_state_handle,
     };
 }

docs/architecture/ecstore-module-split-plan.md

@@ -189,6 +189,10 @@ Required contracts before crate movement:
 - `ReplicationObjectBridge`: app and SetDisks object write/delete replication
   decisions and scheduling are exposed through the contract type in
   `crates/ecstore/src/bucket/replication/replication_object_bridge.rs`.
+- `ObsReplicationStatsSnapshot`: observability reads replication bucket/site
+  metrics through obs-local snapshot DTOs in
+  `crates/obs/src/metrics/storage_api.rs` instead of carrying the ECStore
+  replication stats handle through collectors.
 - `ReplicationScannerBridge`: scanner-originated replication heal scheduling is
   exposed through the contract type in
   `crates/ecstore/src/bucket/replication/replication_scanner_bridge.rs`.

docs/architecture/obs-ecstore-dependency-inventory.md

@@ -21,7 +21,7 @@ until the contracts below are extracted.
 | `rustfs_ecstore::api::bucket::metadata_sys::get_quota_config` | `stats_collector.rs` | Behavior dependency | Reads per-bucket quota limits used in bucket usage metrics. |
 | `rustfs_ecstore::api::bucket::bandwidth::monitor::Monitor` | `runtime_sources.rs`, `stats_collector.rs` | Type and runtime dependency | Reads replication bandwidth reports from the global bucket monitor handle. |
 | `rustfs_ecstore::api::runtime::bucket_monitor` | `runtime_sources.rs` | Runtime dependency | Resolves the global bucket bandwidth monitor for metric collection. |
-| `rustfs_ecstore::api::bucket::replication::{GLOBAL_REPLICATION_STATS, ReplicationStats}` | `runtime_sources.rs`, `stats_collector.rs` | Type and runtime dependency | Reads replication status, transfer, failure, and site-replication stats. |
+| `rustfs_ecstore::api::bucket::replication::get_global_replication_stats` | `storage_api.rs` snapshot helpers | Runtime dependency | Reads replication status, transfer, failure, and site-replication stats, then projects them into obs-local snapshot DTOs. |
 | `rustfs_ecstore::api::bucket::lifecycle::bucket_lifecycle_ops::{GLOBAL_ExpiryState, GLOBAL_TransitionState}` | `runtime_sources.rs`, `stats_collector.rs` | Runtime dependency | Reads lifecycle expiry and transition queue counters. |
 | `rustfs_ecstore::api::error::Result` as `ObsEcstoreResult` | `stats_collector.rs` | Type dependency | Preserves ECStore error propagation while data-usage behavior remains ECStore-owned. |
 
@@ -29,11 +29,10 @@ until the contracts below are extracted.
 
 The remaining coupling is not just a dependency declaration problem:
 
-- type coupling: `ObsStore`, `ObsEcstoreResult`, `ObsReplicationStats`,
-  `ObsBucketBandwidthMonitor`, `StorageAdminApi`, `BucketOperations`, and
-  `BucketOptions`;
+- type coupling: `ObsStore`, `ObsEcstoreResult`, `ObsBucketBandwidthMonitor`,
+  `StorageAdminApi`, `BucketOperations`, and `BucketOptions`;
 - runtime handle coupling: object-store handle, bucket monitor, replication
-  stats, expiry state, and transition state;
+  stats inside snapshot helpers, expiry state, and transition state;
 - behavior coupling: data-usage loading, quota lookup, and capacity math.
 
 Removing `rustfs-ecstore` from `crates/obs/Cargo.toml` is unsafe until those
@@ -43,8 +42,8 @@ three categories have replacement contracts and compile coverage.
 
 1. Keep all direct ECStore and storage-api imports centralized in
    `crates/obs/src/metrics/storage_api.rs`.
-2. Keep projecting ECStore data-usage output into obs-local DTOs before
-   collectors consume it.
+2. Keep projecting ECStore data-usage and replication stats output into
+   obs-local DTOs before collectors consume it.
 3. Introduce obs-owned provider traits for storage info, bucket info, quota,
    data usage, replication, bandwidth, and lifecycle queue snapshots.
 4. Implement those traits in ECStore or an ECStore-owned adapter crate after the
@@ -58,6 +57,8 @@ The architecture guard enforces this inventory boundary:
 
 - `crates/obs/src/metrics/storage_api.rs` is the only `rustfs-obs` source file
   allowed to reference `rustfs_ecstore` or `rustfs_storage_api`;
+- raw replication stats handles and ECStore replication stat methods must stay
+  behind the snapshot helpers in `crates/obs/src/metrics/storage_api.rs`;
 - `rustfs-obs` must not add `storage_compat.rs` or `ecstore_compat.rs`
   passthrough bridges;
 - future extraction PRs must update this inventory and the guard in the same

scripts/check_architecture_migration_rules.sh

@@ -201,6 +201,7 @@ REPLICATION_FACADE_BYPASS_HITS_FILE="${TMP_DIR}/replication_facade_bypass_hits.t
 REPLICATION_FACADE_WILDCARD_EXPORT_HITS_FILE="${TMP_DIR}/replication_facade_wildcard_export_hits.txt"
 ADMIN_REPLICATION_DTO_BOUNDARY_BYPASS_HITS_FILE="${TMP_DIR}/admin_replication_dto_boundary_bypass_hits.txt"
 APP_REPLICATION_DTO_BOUNDARY_BYPASS_HITS_FILE="${TMP_DIR}/app_replication_dto_boundary_bypass_hits.txt"
+OBS_REPLICATION_STATS_BOUNDARY_BYPASS_HITS_FILE="${TMP_DIR}/obs_replication_stats_boundary_bypass_hits.txt"
 REPLICATION_BANDWIDTH_BOUNDARY_BYPASS_HITS_FILE="${TMP_DIR}/replication_bandwidth_boundary_bypass_hits.txt"
 REPLICATION_CONFIG_STORE_BYPASS_HITS_FILE="${TMP_DIR}/replication_config_store_bypass_hits.txt"
 REPLICATION_ERROR_BOUNDARY_BYPASS_HITS_FILE="${TMP_DIR}/replication_error_boundary_bypass_hits.txt"
@@ -2509,6 +2510,24 @@ if [[ -s "$REPLICATION_FACADE_BYPASS_HITS_FILE" ]]; then
   report_failure "replication facade imports must stay in local storage_api boundaries: $(paste -sd '; ' "$REPLICATION_FACADE_BYPASS_HITS_FILE")"
 fi
 
+(
+  cd "$ROOT_DIR"
+  {
+    rg -n --with-filename '\b(ObsReplicationStats|obs_get_global_replication_stats|replication_stats_handle|get_sr_metrics_for_node|get_proxy_stats|mrf_stats)\b' \
+      crates/obs/src/metrics \
+      --glob '*.rs' \
+      --glob '!crates/obs/src/metrics/storage_api.rs' || true
+    rg -n --with-filename 'rustfs_ecstore::api::bucket::replication' \
+      crates/obs/src/metrics \
+      --glob '*.rs' \
+      --glob '!crates/obs/src/metrics/storage_api.rs' || true
+  }
+) >"$OBS_REPLICATION_STATS_BOUNDARY_BYPASS_HITS_FILE"
+
+if [[ -s "$OBS_REPLICATION_STATS_BOUNDARY_BYPASS_HITS_FILE" ]]; then
+  report_failure "obs replication stats access must stay behind crates/obs/src/metrics/storage_api.rs snapshot helpers: $(paste -sd '; ' "$OBS_REPLICATION_STATS_BOUNDARY_BYPASS_HITS_FILE")"
+fi
+
 (
   cd "$ROOT_DIR"
   rg -n --with-filename 'pub(?:\(crate\))?\s+use\s+(?:config|datatypes|replication_pool|replication_resyncer|replication_state|rule)::\*;' \