Use better host names for metric errors (#18188)

Typically hosts would end up like this:

```
   "hosts": [
        ":9000",
        ":9000",
        ":9000",
...
```

Also add host name to errors.
This commit is contained in:
Klaus Post 2023-10-09 17:27:11 -07:00 committed by GitHub
parent 2b4531f069
commit 7cd08594f6
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 7 additions and 20 deletions

View File

@ -20,7 +20,6 @@ package cmd
import ( import (
"context" "context"
"fmt" "fmt"
"net/http"
"time" "time"
"github.com/minio/madmin-go/v3" "github.com/minio/madmin-go/v3"
@ -42,18 +41,6 @@ func collectLocalMetrics(types madmin.MetricType, opts collectMetricsOpts) (m ma
return return
} }
byHostName := globalMinioAddr
if len(opts.hosts) > 0 {
server := getLocalServerProperty(globalEndpoints, &http.Request{
Host: globalLocalNodeName,
})
if _, ok := opts.hosts[server.Endpoint]; ok {
byHostName = server.Endpoint
} else {
return
}
}
if types.Contains(madmin.MetricsDisk) { if types.Contains(madmin.MetricsDisk) {
m.ByDisk = make(map[string]madmin.DiskMetric) m.ByDisk = make(map[string]madmin.DiskMetric)
aggr := madmin.DiskMetric{ aggr := madmin.DiskMetric{
@ -87,7 +74,7 @@ func collectLocalMetrics(types madmin.MetricType, opts collectMetricsOpts) (m ma
} }
netStats, err := net.GetInterfaceNetStats(globalInternodeInterface) netStats, err := net.GetInterfaceNetStats(globalInternodeInterface)
if err != nil { if err != nil {
m.Errors = append(m.Errors, err.Error()) m.Errors = append(m.Errors, fmt.Sprintf("%s: %v (nicstats)", globalMinioAddr, err.Error()))
} else { } else {
m.Aggregated.Net.NetStats = netStats m.Aggregated.Net.NetStats = netStats
} }
@ -104,18 +91,18 @@ func collectLocalMetrics(types madmin.MetricType, opts collectMetricsOpts) (m ma
} }
cm, err := c.Times(false) cm, err := c.Times(false)
if err != nil { if err != nil {
m.Errors = append(m.Errors, err.Error()) m.Errors = append(m.Errors, fmt.Sprintf("%s: %v (cputimes)", globalMinioAddr, err.Error()))
} else { } else {
// not collecting per-cpu stats, so there will be only one element // not collecting per-cpu stats, so there will be only one element
if len(cm) == 1 { if len(cm) == 1 {
m.Aggregated.CPU.TimesStat = &cm[0] m.Aggregated.CPU.TimesStat = &cm[0]
} else { } else {
m.Errors = append(m.Errors, fmt.Sprintf("Expected one CPU stat, got %d", len(cm))) m.Errors = append(m.Errors, fmt.Sprintf("%s: Expected one CPU stat, got %d", globalMinioAddr, len(cm)))
} }
} }
loadStat, err := load.Avg() loadStat, err := load.Avg()
if err != nil { if err != nil {
m.Errors = append(m.Errors, err.Error()) m.Errors = append(m.Errors, fmt.Sprintf("%s: %v (loadStat)", globalMinioAddr, err.Error()))
} else { } else {
m.Aggregated.CPU.LoadStat = loadStat m.Aggregated.CPU.LoadStat = loadStat
} }
@ -123,8 +110,8 @@ func collectLocalMetrics(types madmin.MetricType, opts collectMetricsOpts) (m ma
// Add types... // Add types...
// ByHost is a shallow reference, so careful about sharing. // ByHost is a shallow reference, so careful about sharing.
m.ByHost = map[string]madmin.Metrics{byHostName: m.Aggregated} m.ByHost = map[string]madmin.Metrics{globalMinioAddr: m.Aggregated}
m.Hosts = append(m.Hosts, byHostName) m.Hosts = append(m.Hosts, globalMinioAddr)
return m return m
} }

View File

@ -777,7 +777,7 @@ func (sys *NotificationSys) GetMetrics(ctx context.Context, t madmin.MetricType,
for index, err := range g.Wait() { for index, err := range g.Wait() {
if err != nil { if err != nil {
reply[index].Errors = []string{err.Error()} reply[index].Errors = []string{fmt.Sprintf("%s: %s (rpc)", sys.peerClients[index].String(), err.Error())}
} }
} }
return reply return reply