MyFavMirrors/minio
1
0
Fork 0
mirror of https://github.com/minio/minio.git synced 2025-11-07 21:02:58 -05:00
Code Issues Packages Projects Releases Wiki Activity

metrics: Add replication latency metrics (#13515)

Browse Source 
Add a new Prometheus metric for bucket replication latency

e.g.:
minio_bucket_replication_latency_ns{
    bucket="testbucket",
    operation="upload",
    range="LESS_THAN_1_MiB",
    server="127.0.0.1:9001",
    targetArn="arn:minio:replication::45da043c-14f5-4da4-9316-aba5f77bf730:testbucket"} 2.2015663e+07

Co-authored-by: Klaus Post <klauspost@gmail.com>
This commit is contained in:
Anis Elleuch
2021-11-17 21:10:57 +01:00
committed by GitHub
parent 5b68f8ea6a
commit 4caed7cc0d
11 changed files with 1237 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
cmd/bucket-replication-stats.go
View File

@@ -20,7 +20,6 @@ package cmd
import (
"context"
"sync"
"sync/atomic"
"time"
"github.com/minio/minio/internal/bucket/replication"
@@ -68,16 +67,18 @@ func (r *ReplicationStats) UpdateReplicaStat(bucket string, n int64) {
if !ok {
bs = &BucketReplicationStats{Stats: make(map[string]*BucketReplicationStat)}
}
atomic.AddInt64(&bs.ReplicaSize, n)
bs.ReplicaSize += n
r.Cache[bucket] = bs
}
// Update updates in-memory replication statistics with new values.
func (r *ReplicationStats) Update(bucket string, arn string, n int64, status, prevStatus replication.StatusType, opType replication.Type) {
func (r *ReplicationStats) Update(bucket string, arn string, n int64, duration time.Duration, status, prevStatus replication.StatusType, opType replication.Type) {
if r == nil {
return
}
r.RLock()
r.Lock()
defer r.Unlock()
bs, ok := r.Cache[bucket]
if !ok {
bs = &BucketReplicationStats{Stats: make(map[string]*BucketReplicationStat)}
@@ -86,36 +87,37 @@ func (r *ReplicationStats) Update(bucket string, arn string, n int64, status, pr
if !ok {
b = &BucketReplicationStat{}
}
r.RUnlock()
switch status {
case replication.Completed:
switch prevStatus { // adjust counters based on previous state
case replication.Failed:
atomic.AddInt64(&b.FailedCount, -1)
b.FailedCount--
}
if opType == replication.ObjectReplicationType {
atomic.AddInt64(&b.ReplicatedSize, n)
b.ReplicatedSize += n
switch prevStatus {
case replication.Failed:
atomic.AddInt64(&b.FailedSize, -1*n)
b.FailedSize -= n
}
if duration > 0 {
b.Latency.update(n, duration)
}
}
case replication.Failed:
if opType == replication.ObjectReplicationType {
if prevStatus == replication.Pending {
atomic.AddInt64(&b.FailedSize, n)
atomic.AddInt64(&b.FailedCount, 1)
b.FailedSize += n
b.FailedCount++
}
}
case replication.Replica:
if opType == replication.ObjectReplicationType {
atomic.AddInt64(&b.ReplicaSize, n)
b.ReplicaSize += n
}
}
r.Lock()
bs.Stats[arn] = b
r.Cache[bucket] = bs
r.Unlock()
}
// GetInitialUsage get replication metrics available at the time of cluster initialization