prometheus/scrape/manager.go

300 lines
7.4 KiB
Go
Raw Normal View History

2017-11-25 05:13:54 -08:00
// Copyright 2013 The Prometheus Authors
// 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 scrape
2017-11-25 05:13:54 -08:00
import (
"encoding"
"fmt"
"hash/fnv"
"net"
"os"
2018-01-18 03:49:42 -08:00
"reflect"
2018-01-17 03:46:17 -08:00
"sync"
"time"
2017-11-25 05:13:54 -08:00
"github.com/go-kit/kit/log"
"github.com/go-kit/kit/log/level"
"github.com/pkg/errors"
2017-11-25 05:13:54 -08:00
"github.com/prometheus/prometheus/config"
Refactor SD configuration to remove `config` dependency (#3629) * refactor: move targetGroup struct and CheckOverflow() to their own package * refactor: move auth and security related structs to a utility package, fix import error in utility package * refactor: Azure SD, remove SD struct from config * refactor: DNS SD, remove SD struct from config into dns package * refactor: ec2 SD, move SD struct from config into the ec2 package * refactor: file SD, move SD struct from config to file discovery package * refactor: gce, move SD struct from config to gce discovery package * refactor: move HTTPClientConfig and URL into util/config, fix import error in httputil * refactor: consul, move SD struct from config into consul discovery package * refactor: marathon, move SD struct from config into marathon discovery package * refactor: triton, move SD struct from config to triton discovery package, fix test * refactor: zookeeper, move SD structs from config to zookeeper discovery package * refactor: openstack, remove SD struct from config, move into openstack discovery package * refactor: kubernetes, move SD struct from config into kubernetes discovery package * refactor: notifier, use targetgroup package instead of config * refactor: tests for file, marathon, triton SD - use targetgroup package instead of config.TargetGroup * refactor: retrieval, use targetgroup package instead of config.TargetGroup * refactor: storage, use config util package * refactor: discovery manager, use targetgroup package instead of config.TargetGroup * refactor: use HTTPClient and TLS config from configUtil instead of config * refactor: tests, use targetgroup package instead of config.TargetGroup * refactor: fix tagetgroup.Group pointers that were removed by mistake * refactor: openstack, kubernetes: drop prefixes * refactor: remove import aliases forced due to vscode bug * refactor: move main SD struct out of config into discovery/config * refactor: rename configUtil to config_util * refactor: rename yamlUtil to yaml_config * refactor: kubernetes, remove prefixes * refactor: move the TargetGroup package to discovery/ * refactor: fix order of imports
2017-12-29 12:01:34 -08:00
"github.com/prometheus/prometheus/discovery/targetgroup"
"github.com/prometheus/prometheus/pkg/labels"
2017-11-25 05:13:54 -08:00
"github.com/prometheus/prometheus/storage"
)
// Appendable returns an Appender.
type Appendable interface {
Appender() (storage.Appender, error)
}
// NewManager is the Manager constructor
func NewManager(logger log.Logger, app Appendable) *Manager {
if logger == nil {
logger = log.NewNopLogger()
}
return &Manager{
2017-11-25 05:13:54 -08:00
append: app,
logger: logger,
scrapeConfigs: make(map[string]*config.ScrapeConfig),
scrapePools: make(map[string]*scrapePool),
graceShut: make(chan struct{}),
triggerReload: make(chan struct{}, 1),
2017-11-25 05:13:54 -08:00
}
}
// Manager maintains a set of scrape pools and manages start/stop cycles
2017-11-25 05:13:54 -08:00
// when receiving new target groups form the discovery manager.
type Manager struct {
logger log.Logger
append Appendable
graceShut chan struct{}
jitterSeed uint64 // Global jitterSeed seed is used to spread scrape workload across HA setup.
mtxScrape sync.Mutex // Guards the fields below.
2017-11-25 05:13:54 -08:00
scrapeConfigs map[string]*config.ScrapeConfig
scrapePools map[string]*scrapePool
targetSets map[string][]*targetgroup.Group
triggerReload chan struct{}
2017-11-25 05:13:54 -08:00
}
// Run receives and saves target set updates and triggers the scraping loops reloading.
// Reloading happens in the background so that it doesn't block receiving targets updates.
func (m *Manager) Run(tsets <-chan map[string][]*targetgroup.Group) error {
go m.reloader()
2017-11-25 05:13:54 -08:00
for {
select {
case ts := <-tsets:
m.updateTsets(ts)
select {
case m.triggerReload <- struct{}{}:
default:
}
case <-m.graceShut:
return nil
2017-11-25 05:13:54 -08:00
}
}
}
func (m *Manager) reloader() {
ticker := time.NewTicker(5 * time.Second)
defer ticker.Stop()
for {
select {
case <-m.graceShut:
return
case <-ticker.C:
select {
case <-m.triggerReload:
m.reload()
case <-m.graceShut:
return
}
}
}
}
func (m *Manager) reload() {
m.mtxScrape.Lock()
var wg sync.WaitGroup
for setName, groups := range m.targetSets {
if _, ok := m.scrapePools[setName]; !ok {
scrapeConfig, ok := m.scrapeConfigs[setName]
if !ok {
level.Error(m.logger).Log("msg", "error reloading target set", "err", "invalid config id:"+setName)
continue
}
sp, err := newScrapePool(scrapeConfig, m.append, m.jitterSeed, log.With(m.logger, "scrape_pool", setName))
if err != nil {
level.Error(m.logger).Log("msg", "error creating new scrape pool", "err", err, "scrape_pool", setName)
continue
}
m.scrapePools[setName] = sp
}
wg.Add(1)
// Run the sync in parallel as these take a while and at high load can't catch up.
go func(sp *scrapePool, groups []*targetgroup.Group) {
sp.Sync(groups)
wg.Done()
}(m.scrapePools[setName], groups)
}
m.mtxScrape.Unlock()
wg.Wait()
}
// setJitterSeed calculates a global jitterSeed per server relying on extra label set.
func (m *Manager) setJitterSeed(labels labels.Labels) error {
h := fnv.New64a()
hostname, err := getFqdn()
if err != nil {
return err
}
if _, err := fmt.Fprintf(h, "%s%s", hostname, labels.String()); err != nil {
return err
}
m.jitterSeed = h.Sum64()
return nil
}
// Stop cancels all running scrape pools and blocks until all have exited.
func (m *Manager) Stop() {
m.mtxScrape.Lock()
defer m.mtxScrape.Unlock()
for _, sp := range m.scrapePools {
sp.stop()
}
close(m.graceShut)
}
func (m *Manager) updateTsets(tsets map[string][]*targetgroup.Group) {
m.mtxScrape.Lock()
m.targetSets = tsets
m.mtxScrape.Unlock()
}
2017-11-25 05:13:54 -08:00
// ApplyConfig resets the manager's target providers and job configurations as defined by the new cfg.
func (m *Manager) ApplyConfig(cfg *config.Config) error {
m.mtxScrape.Lock()
defer m.mtxScrape.Unlock()
2018-01-17 03:46:17 -08:00
c := make(map[string]*config.ScrapeConfig)
for _, scfg := range cfg.ScrapeConfigs {
c[scfg.JobName] = scfg
2017-11-25 05:13:54 -08:00
}
2018-01-17 03:46:17 -08:00
m.scrapeConfigs = c
2018-01-18 03:49:42 -08:00
if err := m.setJitterSeed(cfg.GlobalConfig.ExternalLabels); err != nil {
return err
}
// Cleanup and reload pool if the configuration has changed.
var failed bool
2018-01-18 03:49:42 -08:00
for name, sp := range m.scrapePools {
if cfg, ok := m.scrapeConfigs[name]; !ok {
sp.stop()
delete(m.scrapePools, name)
} else if !reflect.DeepEqual(sp.config, cfg) {
err := sp.reload(cfg)
if err != nil {
level.Error(m.logger).Log("msg", "error reloading scrape pool", "err", err, "scrape_pool", name)
failed = true
}
2018-01-18 03:49:42 -08:00
}
}
if failed {
return errors.New("failed to apply the new configuration")
}
2017-11-25 05:13:54 -08:00
return nil
}
// TargetsAll returns active and dropped targets grouped by job_name.
func (m *Manager) TargetsAll() map[string][]*Target {
m.mtxScrape.Lock()
defer m.mtxScrape.Unlock()
2018-01-17 03:46:17 -08:00
targets := make(map[string][]*Target, len(m.scrapePools))
for tset, sp := range m.scrapePools {
targets[tset] = append(sp.ActiveTargets(), sp.DroppedTargets()...)
}
return targets
}
2018-01-17 03:46:17 -08:00
// TargetsActive returns the active targets currently being scraped.
func (m *Manager) TargetsActive() map[string][]*Target {
m.mtxScrape.Lock()
defer m.mtxScrape.Unlock()
2017-11-25 05:13:54 -08:00
var (
wg sync.WaitGroup
mtx sync.Mutex
)
targets := make(map[string][]*Target, len(m.scrapePools))
wg.Add(len(m.scrapePools))
for tset, sp := range m.scrapePools {
// Running in parallel limits the blocking time of scrapePool to scrape
// interval when there's an update from SD.
go func(tset string, sp *scrapePool) {
mtx.Lock()
targets[tset] = sp.ActiveTargets()
mtx.Unlock()
wg.Done()
}(tset, sp)
}
wg.Wait()
return targets
}
// TargetsDropped returns the dropped targets during relabelling.
func (m *Manager) TargetsDropped() map[string][]*Target {
m.mtxScrape.Lock()
defer m.mtxScrape.Unlock()
2017-11-25 05:13:54 -08:00
targets := make(map[string][]*Target, len(m.scrapePools))
for tset, sp := range m.scrapePools {
targets[tset] = sp.DroppedTargets()
2018-01-14 11:42:31 -08:00
}
return targets
2017-11-25 05:13:54 -08:00
}
// getFqdn returns a FQDN if it's possible, otherwise falls back to hostname.
func getFqdn() (string, error) {
hostname, err := os.Hostname()
if err != nil {
return "", err
}
ips, err := net.LookupIP(hostname)
if err != nil {
// Return the system hostname if we can't look up the IP address.
return hostname, nil
}
lookup := func(ipStr encoding.TextMarshaler) (string, error) {
ip, err := ipStr.MarshalText()
if err != nil {
return "", err
}
hosts, err := net.LookupAddr(string(ip))
if err != nil || len(hosts) == 0 {
return "", err
}
return hosts[0], nil
}
for _, addr := range ips {
if ip := addr.To4(); ip != nil {
if fqdn, err := lookup(ip); err == nil {
return fqdn, nil
}
}
if ip := addr.To16(); ip != nil {
if fqdn, err := lookup(ip); err == nil {
return fqdn, nil
}
}
}
return hostname, nil
}