lostakj/pve-exporter
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Metrics now have expiration if not updated

Browse Source
This commit is contained in:
Jan Lošťák
2025-02-22 21:24:03 +01:00
parent 2ed310eef7
commit f78df9d3e3
10 changed files with 455 additions and 232 deletions
Download Patch File Download Diff File Expand all files Collapse all files

147
metrics/pve_node_status_collector.go
View File

@@ -2,9 +2,9 @@ package metrics
import (
"strconv"
"time"
"github.com/prometheus/client_golang/prometheus"
"github.com/prometheus/client_golang/prometheus/promauto"
log "github.com/sirupsen/logrus"
"lostak.dev/pve-exporter/proxmox"
)
@@ -12,212 +12,254 @@ import (
// PVE cluster state collector.
type PveNodeStatusCollector struct {
apiClient *proxmox.PveApiClient // PVE API client instance.
registry *TTLRegistry // TTL metrics registry.
state *prometheus.GaugeVec // Node state prometheus gauge.
uptime *prometheus.GaugeVec // Node uptime in seconds prometheus gauge.
cpus *prometheus.GaugeVec // Node CPU count prometheus gauge.
cpuUsage *prometheus.GaugeVec // Node CPU usage in percent prometheus gauge.
memBytes *prometheus.GaugeVec // Node total RAM capacity in bytes prometheus gauge.
memBytesUsed *prometheus.GaugeVec // Node RAM usage in bytes prometheus gauge.
memBytesFree *prometheus.GaugeVec // Node RAM free in bytes prometheus gauge.
ksmShared *prometheus.GaugeVec // Node Kernel samepage shared in bytes prometheus gauge.
cgroupMode *prometheus.GaugeVec // Node CGroups mode prometheus gauge.
load1 *prometheus.GaugeVec // Node load1 unix like (CPU seconds) prometheus gauge.
load5 *prometheus.GaugeVec // Node load5 unix like (CPU seconds) prometheus gauge.
load15 *prometheus.GaugeVec // Node load15 unix like (CPU seconds) prometheus gauge.
fSFree *prometheus.GaugeVec // Node filesystem free space in bytes prometheus gauge.
fSUsed *prometheus.GaugeVec // Node filesystem used space in bytes prometheus gauge.
fSTotal *prometheus.GaugeVec // Node filesystem total space in bytes prometheus gauge.
fSAvail *prometheus.GaugeVec // Node filesystem available capacity in bytes prometheus gauge.
cpuInfo *prometheus.GaugeVec // Node CPU info prometheus gauge.
systemInfo *prometheus.GaugeVec // Node system info prometheus gauge.
time *prometheus.GaugeVec // Node time prometheus gauge.
localTime *prometheus.GaugeVec // Node localtime prometheus gauge.
state *TTLGaugeVec // Node state prometheus gauge.
uptime *TTLGaugeVec // Node uptime in seconds prometheus gauge.
cpus *TTLGaugeVec // Node CPU count prometheus gauge.
cpuUsage *TTLGaugeVec // Node CPU usage in percent prometheus gauge.
memBytes *TTLGaugeVec // Node total RAM capacity in bytes prometheus gauge.
memBytesUsed *TTLGaugeVec // Node RAM usage in bytes prometheus gauge.
memBytesFree *TTLGaugeVec // Node RAM free in bytes prometheus gauge.
ksmShared *TTLGaugeVec // Node Kernel samepage shared in bytes prometheus gauge.
cgroupMode *TTLGaugeVec // Node CGroups mode prometheus gauge.
load1 *TTLGaugeVec // Node load1 unix like (CPU seconds) prometheus gauge.
load5 *TTLGaugeVec // Node load5 unix like (CPU seconds) prometheus gauge.
load15 *TTLGaugeVec // Node load15 unix like (CPU seconds) prometheus gauge.
fSFree *TTLGaugeVec // Node filesystem free space in bytes prometheus gauge.
fSUsed *TTLGaugeVec // Node filesystem used space in bytes prometheus gauge.
fSTotal *TTLGaugeVec // Node filesystem total space in bytes prometheus gauge.
fSAvail *TTLGaugeVec // Node filesystem available capacity in bytes prometheus gauge.
cpuInfo *TTLGaugeVec // Node CPU info prometheus gauge.
systemInfo *TTLGaugeVec // Node system info prometheus gauge.
time *TTLGaugeVec // Node time prometheus gauge.
localTime *TTLGaugeVec // Node localtime prometheus gauge.
}
// Create new instance of PVE cluster state collector.
func NewPveNodeStatusCollector(apiClient *proxmox.PveApiClient) *PveNodeStatusCollector {
func NewPveNodeStatusCollector(apiClient *proxmox.PveApiClient, registry *TTLRegistry) *PveNodeStatusCollector {
c := PveNodeStatusCollector{apiClient: apiClient}
c.registry = registry
// Node state.
c.state = promauto.NewGaugeVec(
c.state = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_state",
Help: "Node state.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.state)
// Node uptime.
c.uptime = promauto.NewGaugeVec(
c.uptime = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_uptime",
Help: "Node uptime.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.uptime)
// Node cpu count.
c.cpus = promauto.NewGaugeVec(
c.cpus = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_cpu_count",
Help: "Node CPU count.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.cpus)
// Node CPU usage.
c.cpuUsage = promauto.NewGaugeVec(
c.cpuUsage = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_cpu_usage",
Help: "Cluster node CPU usage %.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.cpuUsage)
// Node memory in bytes.
c.memBytes = promauto.NewGaugeVec(
c.memBytes = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_memory_total_bytes",
Help: "Node total memory in bytes.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.memBytes)
// Cluster node memory used in bytes.
c.memBytesUsed = promauto.NewGaugeVec(
c.memBytesUsed = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_memory_used_bytes",
Help: "Node used memory in bytes.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.memBytesUsed)
// Node memory free in bytes.
c.memBytesFree = promauto.NewGaugeVec(
c.memBytesFree = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_memory_free_bytes",
Help: "Node free memory in bytes.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.memBytesFree)
// Kernel samepage shared in bytes.
c.ksmShared = promauto.NewGaugeVec(
c.ksmShared = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_ksm_bytes",
Help: "Node kernel samepage shares in bytes.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.ksmShared)
// Node memory cgroup mode.
c.cgroupMode = promauto.NewGaugeVec(
c.cgroupMode = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_cgroup_mode",
Help: "Node cgroup mode.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.cgroupMode)
// Node load 1.
c.load1 = promauto.NewGaugeVec(
c.load1 = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_load1",
Help: "Node CPU load 1 minute average.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.load1)
// Node load 5.
c.load5 = promauto.NewGaugeVec(
c.load5 = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_load5",
Help: "Node CPU load 5 minutes average.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.load5)
// Cluster node load 15.
c.load15 = promauto.NewGaugeVec(
c.load15 = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_load15",
Help: "Node CPU load 15 minutes average.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.load15)
// Node root FS free bytes.
c.fSFree = promauto.NewGaugeVec(
c.fSFree = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_rootfs_free_bytes",
Help: "Node RootFS free bytes.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.fSFree)
// Node root filesystem used bytes.
c.fSUsed = promauto.NewGaugeVec(
c.fSUsed = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_rootfs_used_bytes",
Help: "Node root filesystem used bytes.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.fSUsed)
// Node root filesystem total bytes.
c.fSTotal = promauto.NewGaugeVec(
c.fSTotal = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_rootfs_total_bytes",
Help: "Node root filesystem total bytes.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.fSTotal)
// Node root filesystem avail bytes.
c.fSAvail = promauto.NewGaugeVec(
c.fSAvail = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_rootfs_avail_bytes",
Help: "Node root filesystem avail bytes.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.fSAvail)
// Node CPU info.
c.cpuInfo = promauto.NewGaugeVec(
c.cpuInfo = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_cpuinfo",
Help: "Node CPU info.",
},
[]string{"cluster", "node", "flags", "cores", "model", "sockets", "cpus", "hvm"},
1*time.Minute,
)
c.registry.Register(c.cpuInfo)
// Node system info metrics.
c.systemInfo = promauto.NewGaugeVec(
c.systemInfo = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_systeminfo",
Help: "Node system info.",
},
[]string{"cluster", "node", "kversion", "pveversion", "machine", "sysname", "release"},
1*time.Minute,
)
c.registry.Register(c.systemInfo)
// Node time info.
c.time = promauto.NewGaugeVec(
c.time = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_time",
Help: "Node time.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.time)
// Node localtime info.
c.localTime = promauto.NewGaugeVec(
c.localTime = NewTTLGaugeVec(
prometheus.GaugeOpts{
Name: "pve_node_localtime",
Help: "Node localtime.",
},
[]string{"cluster", "node"},
1*time.Minute,
)
c.registry.Register(c.localTime)
return &c
}
@@ -229,27 +271,6 @@ func (c *PveNodeStatusCollector) CollectMetrics() error {
return err
}
c.state.Reset()
c.uptime.Reset()
c.cpus.Reset()
c.cpuUsage.Reset()
c.memBytes.Reset()
c.memBytesUsed.Reset()
c.memBytesFree.Reset()
c.ksmShared.Reset()
c.cgroupMode.Reset()
c.load1.Reset()
c.load5.Reset()
c.load15.Reset()
c.fSFree.Reset()
c.fSUsed.Reset()
c.fSTotal.Reset()
c.fSAvail.Reset()
c.cpuInfo.Reset()
c.systemInfo.Reset()
c.time.Reset()
c.localTime.Reset()
for _, node := range cluster.NodeStatuses {
labels := prometheus.Labels{
"cluster": cluster.GetClusterName(),