func newMetrics(r prometheus.Registerer) *metrics {
m := &metrics{}
m.gcDuration = prometheus.NewSummary(prometheus.SummaryOpts{
Name: "alertmanager_silences_gc_duration_seconds",
Help: "Duration of the last silence garbage collection cycle.",
})
m.snapshotDuration = prometheus.NewSummary(prometheus.SummaryOpts{
Name: "alertmanager_silences_snapshot_duration_seconds",
Help: "Duration of the last silence snapshot.",
})
m.queriesTotal = prometheus.NewCounter(prometheus.CounterOpts{
Name: "alertmanager_silences_queries_total",
Help: "How many silence queries were received.",
})
m.queryErrorsTotal = prometheus.NewCounter(prometheus.CounterOpts{
Name: "alertmanager_silences_query_errors_total",
Help: "How many silence received queries did not succeed.",
})
m.queryDuration = prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "alertmanager_silences_query_duration_seconds",
Help: "Duration of silence query evaluation.",
})
if r != nil {
r.MustRegister(
m.gcDuration,
m.snapshotDuration,
m.queriesTotal,
m.queryErrorsTotal,
m.queryDuration,
)
}
return m
}
func newMetrics(r prometheus.Registerer) *metrics {
m := &metrics{}
m.gcDuration = prometheus.NewSummary(prometheus.SummaryOpts{
Name: "alertmanager_nflog_gc_duration_seconds",
Help: "Duration of the last notification log garbage collection cycle.",
})
m.snapshotDuration = prometheus.NewSummary(prometheus.SummaryOpts{
Name: "alertmanager_nflog_snapshot_duration_seconds",
Help: "Duration of the last notification log snapshot.",
})
m.queriesTotal = prometheus.NewCounter(prometheus.CounterOpts{
Name: "alertmanager_nflog_queries_total",
Help: "Number of notification log queries were received.",
})
m.queryErrorsTotal = prometheus.NewCounter(prometheus.CounterOpts{
Name: "alertmanager_nflog_query_errors_total",
Help: "Number notification log received queries that failed.",
})
m.queryDuration = prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "alertmanager_nflog_query_duration_seconds",
Help: "Duration of notification log query evaluation.",
})
if r != nil {
r.MustRegister(
m.gcDuration,
m.snapshotDuration,
m.queriesTotal,
m.queryErrorsTotal,
m.queryDuration,
)
}
return m
}
func TestSensorRecordHistogram(t *testing.T) {
testServer := httptest.NewServer(prometheus.UninstrumentedHandler())
defer testServer.Close()
sensor := &Sensor{
Type: "histogram",
collector: prometheus.NewHistogram(prometheus.HistogramOpts{
Namespace: "telemetry",
Subsystem: "sensors",
Name: "TestSensorRecordHistogram",
Help: "help",
})}
prometheus.MustRegister(sensor.collector)
patt := `telemetry_sensors_TestSensorRecordHistogram_bucket{le="([\.0-9|\+Inf]*)"} ([1-9])`
sensor.record("1.2")
resp := getFromTestServer(t, testServer)
expected := [][]string{{"2.5", "1"}, {"5", "1"}, {"10", "1"}, {"+Inf", "1"}}
if !checkBuckets(resp, patt, expected) {
t.Fatalf("Failed to get match for sensor in response")
}
sensor.record("1.2") // same value should add
resp = getFromTestServer(t, testServer)
expected = [][]string{{"2.5", "2"}, {"5", "2"}, {"10", "2"}, {"+Inf", "2"}}
if !checkBuckets(resp, patt, expected) {
t.Fatalf("Failed to get match for sensor in response")
}
sensor.record("4.5") // overlapping should overlap
resp = getFromTestServer(t, testServer)
expected = [][]string{{"2.5", "2"}, {"5", "3"}, {"10", "3"}, {"+Inf", "3"}}
if !checkBuckets(resp, patt, expected) {
t.Fatalf("Failed to get match for sensor in response")
}
}
func (n *Namespace) NewTimer(name, help string) Timer {
t := &timer{
m: prometheus.NewHistogram(n.newTimerOpts(name, help)),
}
n.addMetric(t)
return t
}
// NewSensors creates new sensors from a raw config
func NewSensors(raw []interface{}) ([]*Sensor, error) {
var sensors []*Sensor
if err := utils.DecodeRaw(raw, &sensors); err != nil {
return nil, fmt.Errorf("Sensor configuration error: %v", err)
}
for _, s := range sensors {
check, err := commands.NewCommand(s.CheckExec, s.Timeout)
if err != nil {
return nil, fmt.Errorf("could not parse check in sensor %s: %s", s.Name, err)
}
check.Name = fmt.Sprintf("%s.sensor", s.Name)
s.checkCmd = check
// the prometheus client lib's API here is baffling... they don't expose
// an interface or embed their Opts type in each of the Opts "subtypes",
// so we can't share the initialization.
switch {
case s.Type == "counter":
s.collector = prometheus.NewCounter(prometheus.CounterOpts{
Namespace: s.Namespace,
Subsystem: s.Subsystem,
Name: s.Name,
Help: s.Help,
})
case s.Type == "gauge":
s.collector = prometheus.NewGauge(prometheus.GaugeOpts{
Namespace: s.Namespace,
Subsystem: s.Subsystem,
Name: s.Name,
Help: s.Help,
})
case s.Type == "histogram":
s.collector = prometheus.NewHistogram(prometheus.HistogramOpts{
Namespace: s.Namespace,
Subsystem: s.Subsystem,
Name: s.Name,
Help: s.Help,
})
case s.Type == "summary":
s.collector = prometheus.NewSummary(prometheus.SummaryOpts{
Namespace: s.Namespace,
Subsystem: s.Subsystem,
Name: s.Name,
Help: s.Help,
})
default:
return nil, fmt.Errorf("invalid sensor type: %s", s.Type)
}
// we're going to unregister before every attempt to register
// so that we can reload config
prometheus.Unregister(s.collector)
if err := prometheus.Register(s.collector); err != nil {
return nil, err
}
}
return sensors, nil
}
func ExampleHistogram() {
temps := prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "pond_temperature_celsius",
Help: "The temperature of the frog pond.", // Sorry, we can't measure how badly it smells.
Buckets: prometheus.LinearBuckets(20, 5, 5), // 5 buckets, each 5 centigrade wide.
})
// Simulate some observations.
for i := 0; i < 1000; i++ {
temps.Observe(30 + math.Floor(120*math.Sin(float64(i)*0.1))/10)
}
// Just for demonstration, let's check the state of the histogram by
// (ab)using its Write method (which is usually only used by Prometheus
// internally).
metric := &dto.Metric{}
temps.Write(metric)
fmt.Println(proto.MarshalTextString(metric))
// Output:
// histogram: <
// sample_count: 1000
// sample_sum: 29969.50000000001
// bucket: <
// cumulative_count: 192
// upper_bound: 20
// >
// bucket: <
// cumulative_count: 366
// upper_bound: 25
// >
// bucket: <
// cumulative_count: 501
// upper_bound: 30
// >
// bucket: <
// cumulative_count: 638
// upper_bound: 35
// >
// bucket: <
// cumulative_count: 816
// upper_bound: 40
// >
// >
}
func newHistogram(opts prometheus.Opts, cutoffs []int64) (prometheus.Histogram, func(int64)) {
buckets := make([]float64, len(cutoffs))
for i := range cutoffs {
buckets[i] = float64(cutoffs[i])
}
hOpts := prometheus.HistogramOpts{
Namespace: opts.Namespace,
Subsystem: opts.Subsystem,
Name: opts.Name,
Help: opts.Help,
ConstLabels: opts.ConstLabels,
Buckets: buckets,
}
m := prometheus.NewHistogram(hOpts)
return m, func(n int64) {
m.Observe(float64(n))
}
}
func main() {
router := routers.NewRouter()
var port string
fmt.Printf("len(os.Args) = %d ", len(os.Args))
if len(os.Args) <= 1 {
fmt.Println("Please speficy the port or use default port 8090")
port = "8090"
} else {
port = os.Args[1]
}
fmt.Printf("port = %s", port)
// histogram
temps := prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "pond_temperature_celsius",
Help: "The temperature of the frog pond.", // Sorry, we can't measure how badly it smells.
Buckets: prometheus.LinearBuckets(20, 5, 5), // 5 buckets, each 5 centigrade wide.
})
// Simulate some observations.
for i := 0; i < 1000; i++ {
temps.Observe(30 + math.Floor(120*math.Sin(float64(i)*0.1))/10)
}
// Just for demonstration, let's check the state of the histogram by
// (ab)using its Write method (which is usually only used by Prometheus
// internally).
metric := &dto.Metric{}
temps.Write(metric)
fmt.Println(proto.MarshalTextString(metric))
router.Handle("/metrics", prometheus.Handler())
//log.Fatal(http.ListenAndServe(":8090", router))
log.Fatal(http.ListenAndServe(":"+port, router))
}
// NewHistogram creates a new configured prometheus histogram object.
func NewHistogram(config HistogramConfig) (spec.Histogram, error) {
newHistogram := &histogram{
HistogramConfig: config,
}
if len(newHistogram.Buckets) == 0 {
return nil, maskAnyf(invalidConfigError, "buckets must not be empty")
}
if newHistogram.Help == "" {
return nil, maskAnyf(invalidConfigError, "help must not be empty")
}
if newHistogram.Name == "" {
return nil, maskAnyf(invalidConfigError, "name must not be empty")
}
newHistogram.ClientHistogram = prometheusclient.NewHistogram(prometheusclient.HistogramOpts{
Buckets: newHistogram.Buckets,
Help: newHistogram.Help,
Name: newHistogram.Name,
})
return newHistogram, nil
}
func init() {
flag.StringVar(&iface, "i", "eth0", "Interface to read packets from")
flag.StringVar(&fname, "r", "", "Filename to read from, overrides -i")
flag.IntVar(&snaplen, "s", 65536, "Number of max bytes to read per packet")
flag.StringVar(&tstype, "t", "", "Type of timestamp to use")
flag.IntVar(&port, "P", 9119, "The port number to listen on ")
flag.BoolVar(&verbose, "v", false, "Enable verbose mode")
}
// the currently max seen response time value
var max = 0.0
// prometheus histogram (https://godoc.org/github.com/prometheus/client_golang/prometheus#Histogram)
var rtHistogram = prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "mongodb_histogram_response_time",
Help: "Response time for MongoDB operations",
Buckets: prometheus.ExponentialBuckets(0.00000001, 2, 10),
})
// prometheus summary (https://godoc.org/github.com/prometheus/client_golang/prometheus#Summary)
var rtSummary = prometheus.NewSummary(prometheus.SummaryOpts{
Name: "mongodb_summary_response_time",
Help: "Response time for MongoDB operations",
})
// prometheus gauge (https://godoc.org/github.com/prometheus/client_golang/prometheus#Gauge)
var rtMax = prometheus.NewGauge(prometheus.GaugeOpts{
Namespace: "ognom",
Name: "mongodb_max_response_time",
Help: "Max response time seen for MongoDB operations in the last 10 seconds",
})
// differentiated via a "service" label.
rpcDurations = prometheus.NewSummaryVec(
prometheus.SummaryOpts{
Name: "rpc_durations_seconds",
Help: "RPC latency distributions.",
Objectives: map[float64]float64{0.5: 0.05, 0.9: 0.01, 0.99: 0.001},
},
[]string{"service"},
)
// The same as above, but now as a histogram, and only for the normal
// distribution. The buckets are targeted to the parameters of the
// normal distribution, with 20 buckets centered on the mean, each
// half-sigma wide.
rpcDurationsHistogram = prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "rpc_durations_histogram_seconds",
Help: "RPC latency distributions.",
Buckets: prometheus.LinearBuckets(*normMean-5**normDomain, .5**normDomain, 20),
})
)
func init() {
// Register the summary and the histogram with Prometheus's default registry.
prometheus.MustRegister(rpcDurations)
prometheus.MustRegister(rpcDurationsHistogram)
}
func main() {
flag.Parse()
start := time.Now()
func newMasterCollector(httpClient *httpClient) prometheus.Collector {
metrics := map[prometheus.Collector]func(metricMap, prometheus.Collector) error{
// CPU/Disk/Mem resources in free/used
gauge("master", "cpus", "Current CPU resources in cluster.", "type"): func(m metricMap, c prometheus.Collector) error {
total, ok := m["master/cpus_total"]
used, ok := m["master/cpus_used"]
if !ok {
return notFoundInMap
}
c.(*prometheus.GaugeVec).WithLabelValues("free").Set(total - used)
c.(*prometheus.GaugeVec).WithLabelValues("used").Set(used)
return nil
},
gauge("master", "cpus_revocable", "Current revocable CPU resources in cluster.", "type"): func(m metricMap, c prometheus.Collector) error {
total, ok := m["master/cpus_revocable_total"]
used, ok := m["master/cpus_revocable_used"]
if !ok {
return notFoundInMap
}
c.(*prometheus.GaugeVec).WithLabelValues("free").Set(total - used)
c.(*prometheus.GaugeVec).WithLabelValues("used").Set(used)
return nil
},
gauge("master", "mem", "Current memory resources in cluster.", "type"): func(m metricMap, c prometheus.Collector) error {
total, ok := m["master/mem_total"]
used, ok := m["master/mem_used"]
if !ok {
return notFoundInMap
}
c.(*prometheus.GaugeVec).WithLabelValues("free").Set(total - used)
c.(*prometheus.GaugeVec).WithLabelValues("used").Set(used)
return nil
},
gauge("master", "mem_revocable", "Current revocable memory resources in cluster.", "type"): func(m metricMap, c prometheus.Collector) error {
total, ok := m["master/mem_revocable_total"]
used, ok := m["master/mem_revocable_used"]
if !ok {
return notFoundInMap
}
c.(*prometheus.GaugeVec).WithLabelValues("free").Set(total - used)
c.(*prometheus.GaugeVec).WithLabelValues("used").Set(used)
return nil
},
gauge("master", "disk", "Current disk resources in cluster.", "type"): func(m metricMap, c prometheus.Collector) error {
total, ok := m["master/disk_total"]
used, ok := m["master/disk_used"]
if !ok {
return notFoundInMap
}
c.(*prometheus.GaugeVec).WithLabelValues("free").Set(total - used)
c.(*prometheus.GaugeVec).WithLabelValues("used").Set(used)
return nil
},
gauge("master", "disk_revocable", "Current disk resources in cluster.", "type"): func(m metricMap, c prometheus.Collector) error {
total, ok := m["master/disk_revocable_total"]
used, ok := m["master/disk_revocable_used"]
if !ok {
return notFoundInMap
}
c.(*prometheus.GaugeVec).WithLabelValues("free").Set(total - used)
c.(*prometheus.GaugeVec).WithLabelValues("used").Set(used)
return nil
},
// Master stats about uptime and election state
prometheus.NewGauge(prometheus.GaugeOpts{
Namespace: "mesos",
Subsystem: "master",
Name: "elected",
Help: "1 if master is elected leader, 0 if not",
}): func(m metricMap, c prometheus.Collector) error {
elected, ok := m["master/elected"]
if !ok {
return notFoundInMap
}
c.(prometheus.Gauge).Set(elected)
return nil
},
prometheus.NewGauge(prometheus.GaugeOpts{
Namespace: "mesos",
Subsystem: "master",
Name: "uptime_seconds",
Help: "Number of seconds the master process is running.",
}): func(m metricMap, c prometheus.Collector) error {
uptime, ok := m["master/uptime_secs"]
if !ok {
return notFoundInMap
}
c.(prometheus.Gauge).Set(uptime)
return nil
},
// Master stats about agents
counter("master", "slave_registration_events_total", "Total number of registration events on this master since it booted.", "event"): func(m metricMap, c prometheus.Collector) error {
registrations, ok := m["master/slave_registrations"]
reregistrations, ok := m["master/slave_reregistrations"]
if !ok {
return notFoundInMap
}
c.(*settableCounterVec).Set(registrations, "register")
c.(*settableCounterVec).Set(reregistrations, "reregister")
return nil
},
counter("master", "slave_removal_events_total", "Total number of removal events on this master since it booted.", "event"): func(m metricMap, c prometheus.Collector) error {
scheduled, ok := m["master/slave_shutdowns_scheduled"]
canceled, ok := m["master/slave_shutdowns_canceled"]
completed, ok := m["master/slave_shutdowns_completed"]
removals, ok := m["master/slave_removals"]
if !ok {
return notFoundInMap
}
c.(*settableCounterVec).Set(scheduled, "scheduled")
c.(*settableCounterVec).Set(canceled, "canceled")
c.(*settableCounterVec).Set(completed, "completed")
c.(*settableCounterVec).Set(removals-completed, "died")
return nil
},
gauge("master", "slaves_state", "Current number of slaves known to the master per connection and registration state.", "connection_state", "registration_state"): func(m metricMap, c prometheus.Collector) error {
active, ok := m["master/slaves_active"]
inactive, ok := m["master/slaves_inactive"]
disconnected, ok := m["master/slaves_disconnected"]
if !ok {
return notFoundInMap
}
// FIXME: Make sure those assumptions are right
// Every "active" node is connected to the master
c.(*prometheus.GaugeVec).WithLabelValues("connected", "active").Set(active)
// Every "inactive" node is connected but node sending offers
c.(*prometheus.GaugeVec).WithLabelValues("connected", "inactive").Set(inactive)
// Every "disconnected" node is "inactive"
c.(*prometheus.GaugeVec).WithLabelValues("disconnected", "inactive").Set(disconnected)
// Every "connected" node is either active or inactive
return nil
},
// Master stats about frameworks
gauge("master", "frameworks_state", "Current number of frames known to the master per connection and registration state.", "connection_state", "registration_state"): func(m metricMap, c prometheus.Collector) error {
active, ok := m["master/frameworks_active"]
inactive, ok := m["master/frameworks_inactive"]
disconnected, ok := m["master/frameworks_disconnected"]
if !ok {
return notFoundInMap
}
// FIXME: Make sure those assumptions are right
// Every "active" framework is connected to the master
c.(*prometheus.GaugeVec).WithLabelValues("connected", "active").Set(active)
// Every "inactive" framework is connected but framework sending offers
c.(*prometheus.GaugeVec).WithLabelValues("connected", "inactive").Set(inactive)
// Every "disconnected" framework is "inactive"
c.(*prometheus.GaugeVec).WithLabelValues("disconnected", "inactive").Set(disconnected)
// Every "connected" framework is either active or inactive
return nil
},
prometheus.NewGauge(prometheus.GaugeOpts{
Namespace: "mesos",
Subsystem: "master",
Name: "offers_pending",
Help: "Current number of offers made by the master which aren't yet accepted or declined by frameworks.",
}): func(m metricMap, c prometheus.Collector) error {
offers, ok := m["master/outstanding_offers"]
if !ok {
return notFoundInMap
}
c.(prometheus.Gauge).Set(offers)
// c.(*prometheus.Gauge).Set(offers)
return nil
},
// Master stats about tasks
counter("master", "task_states_exit_total", "Total number of tasks processed by exit state.", "state"): func(m metricMap, c prometheus.Collector) error {
errored, ok := m["master/tasks_error"]
failed, ok := m["master/tasks_failed"]
finished, ok := m["master/tasks_finished"]
killed, ok := m["master/tasks_killed"]
lost, ok := m["master/tasks_lost"]
if !ok {
return notFoundInMap
}
c.(*settableCounterVec).Set(errored, "errored")
c.(*settableCounterVec).Set(failed, "failed")
c.(*settableCounterVec).Set(finished, "finished")
c.(*settableCounterVec).Set(killed, "killed")
c.(*settableCounterVec).Set(lost, "lost")
return nil
},
counter("master", "task_states_current", "Current number of tasks by state.", "state"): func(m metricMap, c prometheus.Collector) error {
running, ok := m["master/tasks_running"]
staging, ok := m["master/tasks_staging"]
starting, ok := m["master/tasks_starting"]
if !ok {
return notFoundInMap
}
c.(*settableCounterVec).Set(running, "running")
c.(*settableCounterVec).Set(staging, "staging")
c.(*settableCounterVec).Set(starting, "starting")
return nil
},
// Master stats about messages
counter("master", "messages_outcomes_total",
"Total number of messages by outcome of operation and direction.",
"source", "destination", "type", "outcome"): func(m metricMap, c prometheus.Collector) error {
frameworkToExecutorValid, ok := m["master/valid_framework_to_executor_messages"]
frameworkToExecutorInvalid, ok := m["master/invalid_framework_to_executor_messages"]
executorToFrameworkValid, ok := m["master/valid_executor_to_framework_messages"]
executorToFrameworkInvalid, ok := m["master/invalid_executor_to_framework_messages"]
// status updates are sent from framework?(FIXME) to slave
// status update acks are sent from slave to framework?
statusUpdateAckValid, ok := m["master/valid_status_update_acknowledgements"]
statusUpdateAckInvalid, ok := m["master/invalid_status_update_acknowledgements"]
statusUpdateValid, ok := m["master/valid_status_updates"]
statusUpdateInvalid, ok := m["master/invalid_status_updates"]
if !ok {
return notFoundInMap
}
c.(*settableCounterVec).Set(frameworkToExecutorValid, "framework", "executor", "", "valid")
c.(*settableCounterVec).Set(frameworkToExecutorInvalid, "framework", "executor", "", "invalid")
c.(*settableCounterVec).Set(executorToFrameworkValid, "executor", "framework", "", "valid")
c.(*settableCounterVec).Set(executorToFrameworkInvalid, "executor", "framework", "", "invalid")
// We consider a ack message simply as a message from slave to framework
c.(*settableCounterVec).Set(statusUpdateValid, "framework", "slave", "status_update", "valid")
c.(*settableCounterVec).Set(statusUpdateInvalid, "framework", "slave", "status_update", "invalid")
c.(*settableCounterVec).Set(statusUpdateAckValid, "slave", "framework", "status_update", "valid")
c.(*settableCounterVec).Set(statusUpdateAckInvalid, "slave", "framework", "status_update", "invalid")
return nil
},
counter("master", "messages_type_total", "Total number of valid messages by type.", "type"): func(m metricMap, c prometheus.Collector) error {
for k, v := range m {
i := strings.Index("master/messages_", k)
if i == -1 {
continue
}
// FIXME: We expose things like messages_framework_to_executor twice
c.(*settableCounterVec).Set(v, k[i:])
}
return nil
},
// Master stats about events
gauge("master", "event_queue_length", "Current number of elements in event queue by type", "type"): func(m metricMap, c prometheus.Collector) error {
dispatches, ok := m["master/event_queue_dispatches"]
httpRequests, ok := m["master/event_queue_http_requests"]
messages, ok := m["master/event_queue_messages"]
if !ok {
return notFoundInMap
}
c.(*prometheus.GaugeVec).WithLabelValues("message").Set(messages)
c.(*prometheus.GaugeVec).WithLabelValues("http_request").Set(httpRequests)
c.(*prometheus.GaugeVec).WithLabelValues("dispatches").Set(dispatches)
return nil
},
// Master stats about registrar
prometheus.NewHistogram(prometheus.HistogramOpts{
Namespace: "mesos",
Subsystem: "master",
Name: "state_store_seconds",
Help: "Registry write latency in seconds",
}): func(m metricMap, c prometheus.Collector) error {
// c.(*prometheus.Histogram).Buckets //FIXME
return nil
},
}
return newMetricCollector(httpClient, metrics)
}
refreshHighWatermarks = "refresh_high_watermarks"
renderView = "render_view"
cutOff = "recency_threshold"
processorName = "processor"
)
var (
diskLatencyHistogram = &prometheus.HistogramSpecification{
Starts: prometheus.LogarithmicSizedBucketsFor(0, 5000),
BucketBuilder: prometheus.AccumulatingBucketBuilder(prometheus.EvictAndReplaceWith(10, prometheus.AverageReducer), 100),
ReportablePercentiles: []float64{0.01, 0.05, 0.5, 0.90, 0.99},
}
curationDuration = prometheus.NewCounter()
curationDurations = prometheus.NewHistogram(diskLatencyHistogram)
curationFilterOperations = prometheus.NewCounter()
storageOperations = prometheus.NewCounter()
storageOperationDurations = prometheus.NewCounter()
storageLatency = prometheus.NewHistogram(diskLatencyHistogram)
queueSizes = prometheus.NewGauge()
storedSamplesCount = prometheus.NewCounter()
)
func recordOutcome(duration time.Duration, err error, success, failure map[string]string) {
labels := success
if err != nil {
labels = failure
}
storageOperations.Increment(labels)
alive = "alive"
failure = "failure"
outcome = "outcome"
state = "state"
success = "success"
unreachable = "unreachable"
)
var (
networkLatencyHistogram = &prometheus.HistogramSpecification{
Starts: prometheus.LogarithmicSizedBucketsFor(0, 1000),
BucketBuilder: prometheus.AccumulatingBucketBuilder(prometheus.EvictAndReplaceWith(10, prometheus.AverageReducer), 100),
ReportablePercentiles: []float64{0.01, 0.05, 0.5, 0.90, 0.99},
}
targetOperationLatencies = prometheus.NewHistogram(networkLatencyHistogram)
retrievalDurations = prometheus.NewHistogram(&prometheus.HistogramSpecification{
Starts: prometheus.LogarithmicSizedBucketsFor(0, 10000),
BucketBuilder: prometheus.AccumulatingBucketBuilder(prometheus.EvictAndReplaceWith(10, prometheus.AverageReducer), 100),
ReportablePercentiles: []float64{0.01, 0.05, 0.5, 0.90, 0.99}})
targetOperations = prometheus.NewCounter()
dnsSDLookupsCount = prometheus.NewCounter()
)
func recordOutcome(err error) {
message := success
if err != nil {
message = failure
}
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package wal
import "github.com/prometheus/client_golang/prometheus"
var (
syncDurations = prometheus.NewHistogram(prometheus.HistogramOpts{
Namespace: "etcd",
Subsystem: "disk",
Name: "wal_fsync_duration_seconds",
Help: "The latency distributions of fsync called by wal.",
Buckets: prometheus.ExponentialBuckets(0.001, 2, 14),
})
)
func init() {
prometheus.MustRegister(syncDurations)
}
Buckets: prometheus.ExponentialBuckets(0.0005, 2, 18),
}, []string{"type"})
sendReqHistogram = prometheus.NewHistogramVec(
prometheus.HistogramOpts{
Namespace: "tidb",
Subsystem: "tikvclient",
Name: "request_seconds",
Help: "Bucketed histogram of sending request duration.",
Buckets: prometheus.ExponentialBuckets(0.0005, 2, 18),
}, []string{"type"})
copBuildTaskHistogram = prometheus.NewHistogram(
prometheus.HistogramOpts{
Namespace: "tidb",
Subsystem: "tikvclient",
Name: "cop_buildtask_seconds",
Help: "Coprocessor buildTask cost time.",
})
copTaskLenHistogram = prometheus.NewHistogram(
prometheus.HistogramOpts{
Namespace: "tidb",
Subsystem: "tikvclient",
Name: "cop_task_len",
Help: "Coprocessor task length.",
Buckets: prometheus.ExponentialBuckets(1, 2, 11),
})
coprocessorCounter = prometheus.NewCounterVec(
prometheus.CounterOpts{
package etcdserver
import (
"time"
"github.com/coreos/etcd/pkg/runtime"
"github.com/prometheus/client_golang/prometheus"
)
var (
// TODO: with label in v3?
proposeDurations = prometheus.NewHistogram(prometheus.HistogramOpts{
Namespace: "etcd",
Subsystem: "server",
Name: "proposal_durations_seconds",
Help: "The latency distributions of committing proposal.",
Buckets: prometheus.ExponentialBuckets(0.001, 2, 14),
})
proposePending = prometheus.NewGauge(prometheus.GaugeOpts{
Namespace: "etcd",
Subsystem: "server",
Name: "pending_proposal_total",
Help: "The total number of pending proposals.",
})
// This is number of proposal failed in client's view.
// The proposal might be later got committed in raft.
proposeFailed = prometheus.NewCounter(prometheus.CounterOpts{
Namespace: "etcd",
Subsystem: "server",
Name: "proposal_failed_total",
var (
queryResultsPath = flag.String("query_results_path",
"/tmp/qr/",
"Path where query results files (page_0.json etc.) are stored")
perPackagePathRe = regexp.MustCompile(`^/perpackage-results/([^/]+)/` +
strconv.Itoa(resultsPerPackage) + `/page_([0-9]+).json$`)
queryDurations = prometheus.NewHistogram(
prometheus.HistogramOpts{
Name: "query_durations_ms",
Help: "Duration of a query in milliseconds.",
Buckets: []float64{
1, 2, 3, 4, 5, 6, 7, 8, 9, 10,
15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100,
150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000,
2000, 3000, 4000, 5000, 6000, 7000, 8000, 9000, 10000,
20000, 30000, 40000, 50000, 60000, 70000, 80000, 90000, 100000,
500000, 1000000,
},
})
)
const (
// NB: All of these constants needs to match those in static/instant.js.
packagesPerPage = 5
resultsPerPackage = 2
resultsPerPage = 10
)
Name: "request_total",
Help: "Counter of tikv-server requests.",
}, []string{"type"})
sendReqHistogram = prometheus.NewHistogramVec(
prometheus.HistogramOpts{
Namespace: "tidb",
Subsystem: "tikvclient",
Name: "request_seconds",
Help: "Bucketed histogram of sending request duration.",
}, []string{"type"})
copBuildTaskHistogram = prometheus.NewHistogram(
prometheus.HistogramOpts{
Namespace: "tidb",
Subsystem: "tikvclient",
Name: "cop_buildtask_seconds",
Help: "Coprocessor buildTask cost time.",
})
copTaskLenHistogram = prometheus.NewHistogram(
prometheus.HistogramOpts{
Namespace: "tidb",
Subsystem: "tikvclient",
Name: "cop_task_len",
Help: "Coprocessor task length.",
Buckets: prometheus.ExponentialBuckets(1, 2, 11),
})
coprocessorCounter = prometheus.NewCounterVec(
prometheus.CounterOpts{
// See the License for the specific language governing permissions and
// limitations under the License.
package metrics
import (
"time"
"github.com/prometheus/client_golang/prometheus"
)
var (
queryMetric = prometheus.NewHistogram(
prometheus.HistogramOpts{
Namespace: "tidb",
Subsystem: "query",
Name: "handle_query_duration_seconds",
Help: "Bucketed histogram of processing time (s) of handled queries.",
Buckets: prometheus.ExponentialBuckets(0.0005, 2, 13),
})
)
// Query is used for add query cost time into metrics.
func Query(costTime time.Duration) {
queryMetric.Observe(float64(costTime))
}
func init() {
prometheus.MustRegister(queryMetric)
}
Namespace: Namespace,
Subsystem: "engine",
Name: "task_failure_count_total",
Help: "Counter of engine schedule task failures.",
}, []string{"type"})
engineReconcileCount = prometheus.NewCounter(prometheus.CounterOpts{
Namespace: Namespace,
Subsystem: "engine",
Name: "reconcile_count_total",
Help: "Counter of reconcile rounds.",
})
engineReconcileDuration = prometheus.NewHistogram(prometheus.HistogramOpts{
Namespace: Namespace,
Subsystem: "engine",
Name: "reconcile_duration_second",
Help: "Histogram of time (in seconds) each schedule round takes.",
})
engineReconcileFailureCount = prometheus.NewCounterVec(prometheus.CounterOpts{
Namespace: Namespace,
Subsystem: "engine",
Name: "reconcile_failure_count_total",
Help: "Counter of scheduling failures.",
}, []string{"type"})
registryOpCount = prometheus.NewCounterVec(prometheus.CounterOpts{
Namespace: Namespace,
Subsystem: "registry",
Name: "operation_count_total",
Help: "Counter of registry operations.",
)
const (
checkInterval = 5 * time.Minute
refreshLockDuration = time.Minute * 2
lockDuration = time.Minute*8 + refreshLockDuration
maxBackOff = 15 * time.Minute
)
var (
log = capnslog.NewPackageLogger("github.com/coreos/clair", "notifier")
notifiers = make(map[string]Notifier)
promNotifierLatencyMilliseconds = prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "clair_notifier_latency_milliseconds",
Help: "Time it takes to send a notification after it's been created.",
})
promNotifierBackendErrorsTotal = prometheus.NewCounterVec(prometheus.CounterOpts{
Name: "clair_notifier_backend_errors_total",
Help: "Number of errors that notifier backends generated.",
}, []string{"backend"})
)
// Notifier represents anything that can transmit notifications.
type Notifier interface {
// Configure attempts to initialize the notifier with the provided configuration.
// It returns whether the notifier is enabled or not.
Configure(*config.NotifierConfig) (bool, error)
// Send informs the existence of the specified notification.
Send(notification database.VulnerabilityNotification) error
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// See the License for the specific language governing permissions and
// limitations under the License.
package tidb
import (
"github.com/prometheus/client_golang/prometheus"
)
var (
sessionExecuteParseDuration = prometheus.NewHistogram(
prometheus.HistogramOpts{
Namespace: "tidb",
Subsystem: "server",
Name: "session_execute_parse_duration",
Help: "Bucketed histogram of processing time (s) in parse SQL.",
Buckets: prometheus.LinearBuckets(0.00004, 0.00001, 13),
})
sessionExecuteCompileDuration = prometheus.NewHistogram(
prometheus.HistogramOpts{
Namespace: "tidb",
Subsystem: "server",
Name: "session_execute_compile_duration",
Help: "Bucketed histogram of processing time (s) in query optimize.",
Buckets: prometheus.LinearBuckets(0.00004, 0.00001, 13),
})
sessionExecuteRunDuration = prometheus.NewHistogram(
prometheus.HistogramOpts{
Namespace: "tidb",
Subsystem: "server",
uniqueLurkingAgentCount = prometheus.NewGauge(prometheus.GaugeOpts{
Name: "unique_lurking_agents",
Subsystem: "presence",
Help: "Number of unique, lurking agents in the presence table.",
})
uniqueWebAgentCount = prometheus.NewGauge(prometheus.GaugeOpts{
Name: "unique_web_agents",
Subsystem: "presence",
Help: "Number of unique agents using the web client in the presence table.",
})
sessionsPerAgent = prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "sessions_per_agent",
Subsystem: "presence",
Help: "Number of simultaneous live sessions for each active agent.",
Buckets: prometheus.LinearBuckets(0, 1, 10),
})
)
func init() {
prometheus.MustRegister(rowCount)
prometheus.MustRegister(activeRowCount)
prometheus.MustRegister(activeRowCountPerRoom)
prometheus.MustRegister(lurkingRowCount)
prometheus.MustRegister(lurkingRowCountPerRoom)
prometheus.MustRegister(uniqueAgentCount)
prometheus.MustRegister(uniqueLurkingAgentCount)
prometheus.MustRegister(uniqueWebAgentCount)
prometheus.MustRegister(sessionsPerAgent)
}
"time"
"github.com/prometheus/client_golang/prometheus"
)
var (
port = flag.String("port", "8000", "Port to listen on.")
)
var (
requests = prometheus.NewCounter(prometheus.CounterOpts{
Name: "requests", Help: "total requests received"})
errors = prometheus.NewCounterVec(prometheus.CounterOpts{
Name: "errors", Help: "total errors served"}, []string{"code"})
latency_ms = prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "latency_ms",
Help: "request latency in milliseconds",
Buckets: prometheus.ExponentialBuckets(1, 2, 20)})
backend_latency_ms = prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "backend_latency_ms",
Help: "request latency in milliseconds",
Buckets: prometheus.ExponentialBuckets(1, 2, 20)})
)
func init() {
prometheus.MustRegister(requests)
prometheus.MustRegister(errors)
prometheus.MustRegister(latency_ms)
prometheus.MustRegister(backend_latency_ms)
}
var (
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package snap
import "github.com/prometheus/client_golang/prometheus"
var (
// TODO: save_fsync latency?
saveDurations = prometheus.NewHistogram(prometheus.HistogramOpts{
Namespace: "etcd_debugging",
Subsystem: "snap",
Name: "save_total_durations_seconds",
Help: "The total latency distributions of save called by snapshot.",
Buckets: prometheus.ExponentialBuckets(0.001, 2, 14),
})
marshallingDurations = prometheus.NewHistogram(prometheus.HistogramOpts{
Namespace: "etcd_debugging",
Subsystem: "snap",
Name: "save_marshalling_durations_seconds",
Help: "The marshalling cost distributions of save called by snapshot.",
Buckets: prometheus.ExponentialBuckets(0.001, 2, 14),
})
)
func init() {
prometheus.MustRegister(saveDurations)
}, []string{"volume"})
promFileSyncs = prometheus.NewCounterVec(prometheus.CounterOpts{
Name: "torus_server_file_syncs",
Help: "Number of times a file has been synced on this server",
}, []string{"volume"})
promFileChangedSyncs = prometheus.NewCounterVec(prometheus.CounterOpts{
Name: "torus_server_file_changed_syncs",
Help: "Number of times a file has been synced on this server, and the file has changed underneath it",
}, []string{"volume"})
promFileWrittenBytes = prometheus.NewCounterVec(prometheus.CounterOpts{
Name: "torus_server_file_written_bytes",
Help: "Number of bytes written to a file on this server",
}, []string{"volume"})
promFileBlockRead = prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "torus_server_file_block_read_us",
Help: "Histogram of ms taken to read a block through the layers and into the file abstraction",
Buckets: prometheus.ExponentialBuckets(50.0, 2, 20),
})
promFileBlockWrite = prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "torus_server_file_block_write_us",
Help: "Histogram of ms taken to write a block through the layers and into the file abstraction",
Buckets: prometheus.ExponentialBuckets(50.0, 2, 20),
})
)
func init() {
prometheus.MustRegister(promOpenINodes)
prometheus.MustRegister(promOpenFiles)
prometheus.MustRegister(promFileSyncs)
prometheus.MustRegister(promFileChangedSyncs)
prometheus.MustRegister(promFileWrittenBytes)
Help: "Total number of events sent by this member.",
})
pendingEventsGauge = prometheus.NewGauge(
prometheus.GaugeOpts{
Namespace: "etcd",
Subsystem: "storage",
Name: "pending_events_total",
Help: "Total number of pending events to be sent.",
})
indexCompactionPauseDurations = prometheus.NewHistogram(
prometheus.HistogramOpts{
Namespace: "etcd",
Subsystem: "storage",
Name: "index_compaction_pause_duration_milliseconds",
Help: "Bucketed histogram of index compaction pause duration.",
// 0.5ms -> 1second
Buckets: prometheus.ExponentialBuckets(0.5, 2, 12),
})
dbCompactionPauseDurations = prometheus.NewHistogram(
prometheus.HistogramOpts{
Namespace: "etcd",
Subsystem: "storage",
Name: "db_compaction_pause_duration_milliseconds",
Help: "Bucketed histogram of db compaction pause duration.",
// 1ms -> 4second
Buckets: prometheus.ExponentialBuckets(1, 2, 13),
})
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package backend
import "github.com/prometheus/client_golang/prometheus"
var (
commitDurations = prometheus.NewHistogram(prometheus.HistogramOpts{
Namespace: "etcd",
Subsystem: "disk",
Name: "backend_commit_duration_seconds",
Help: "The latency distributions of commit called by backend.",
Buckets: prometheus.ExponentialBuckets(0.001, 2, 14),
})
)
func init() {
prometheus.MustRegister(commitDurations)
}
import (
"sync"
"time"
"github.com/prometheus/client_golang/prometheus"
)
const schedulerSubsystem = "scheduler"
var BindingSaturationReportInterval = 1 * time.Second
var (
E2eSchedulingLatency = prometheus.NewHistogram(
prometheus.HistogramOpts{
Subsystem: schedulerSubsystem,
Name: "e2e_scheduling_latency_microseconds",
Help: "E2e scheduling latency (scheduling algorithm + binding)",
Buckets: prometheus.ExponentialBuckets(1000, 2, 15),
},
)
SchedulingAlgorithmLatency = prometheus.NewHistogram(
prometheus.HistogramOpts{
Subsystem: schedulerSubsystem,
Name: "scheduling_algorithm_latency_microseconds",
Help: "Scheduling algorithm latency",
Buckets: prometheus.ExponentialBuckets(1000, 2, 15),
},
)
BindingLatency = prometheus.NewHistogram(
prometheus.HistogramOpts{
Subsystem: schedulerSubsystem,
Name: "binding_latency_microseconds",
