cc-metric-collector/sinks/influxAsyncSink.go
2022-04-02 16:06:03 +02:00

214 lines
6.2 KiB
Go

package sinks
import (
"context"
"crypto/tls"
"encoding/json"
"errors"
"fmt"
"time"
cclog "github.com/ClusterCockpit/cc-metric-collector/internal/ccLogger"
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
stats "github.com/ClusterCockpit/cc-metric-collector/internal/metricRouter"
influxdb2 "github.com/influxdata/influxdb-client-go/v2"
influxdb2Api "github.com/influxdata/influxdb-client-go/v2/api"
)
type InfluxAsyncSinkConfig struct {
defaultSinkConfig
Host string `json:"host,omitempty"`
Port string `json:"port,omitempty"`
Database string `json:"database,omitempty"`
User string `json:"user,omitempty"`
Password string `json:"password,omitempty"`
Organization string `json:"organization,omitempty"`
SSL bool `json:"ssl,omitempty"`
RetentionPol string `json:"retention_policy,omitempty"`
// Maximum number of points sent to server in single request. Default 5000
BatchSize uint `json:"batch_size,omitempty"`
// Interval, in ms, in which is buffer flushed if it has not been already written (by reaching batch size) . Default 1000ms
FlushInterval uint `json:"flush_interval,omitempty"`
InfluxRetryInterval string `json:"retry_interval,omitempty"`
InfluxExponentialBase uint `json:"retry_exponential_base,omitempty"`
InfluxMaxRetries uint `json:"max_retries,omitempty"`
InfluxMaxRetryTime string `json:"max_retry_time,omitempty"`
}
type InfluxAsyncSink struct {
sink
client influxdb2.Client
writeApi influxdb2Api.WriteAPI
errors <-chan error
config InfluxAsyncSinkConfig
influxRetryInterval uint
influxMaxRetryTime uint
sentMetrics int64
statsFlushes int64
statsErrors int64
}
func (s *InfluxAsyncSink) connect() error {
var auth string
var uri string
if s.config.SSL {
uri = fmt.Sprintf("https://%s:%s", s.config.Host, s.config.Port)
} else {
uri = fmt.Sprintf("http://%s:%s", s.config.Host, s.config.Port)
}
if len(s.config.User) == 0 {
auth = s.config.Password
} else {
auth = fmt.Sprintf("%s:%s", s.config.User, s.config.Password)
}
cclog.ComponentDebug(s.name, "Using URI", uri, "Org", s.config.Organization, "Bucket", s.config.Database)
clientOptions := influxdb2.DefaultOptions()
if s.config.BatchSize != 0 {
cclog.ComponentDebug(s.name, "Batch size", s.config.BatchSize)
clientOptions.SetBatchSize(s.config.BatchSize)
}
if s.config.FlushInterval != 0 {
cclog.ComponentDebug(s.name, "Flush interval", s.config.FlushInterval)
clientOptions.SetFlushInterval(s.config.FlushInterval)
}
if s.influxRetryInterval != 0 {
cclog.ComponentDebug(s.name, "MaxRetryInterval", s.influxRetryInterval)
clientOptions.SetMaxRetryInterval(s.influxRetryInterval)
}
if s.influxMaxRetryTime != 0 {
cclog.ComponentDebug(s.name, "MaxRetryTime", s.influxMaxRetryTime)
clientOptions.SetMaxRetryTime(s.influxMaxRetryTime)
}
if s.config.InfluxExponentialBase != 0 {
cclog.ComponentDebug(s.name, "Exponential Base", s.config.InfluxExponentialBase)
clientOptions.SetExponentialBase(s.config.InfluxExponentialBase)
}
if s.config.InfluxMaxRetries != 0 {
cclog.ComponentDebug(s.name, "Max Retries", s.config.InfluxMaxRetries)
clientOptions.SetMaxRetries(s.config.InfluxMaxRetries)
}
clientOptions.SetTLSConfig(
&tls.Config{
InsecureSkipVerify: true,
},
).SetPrecision(time.Second)
s.client = influxdb2.NewClientWithOptions(uri, auth, clientOptions)
s.writeApi = s.client.WriteAPI(s.config.Organization, s.config.Database)
ok, err := s.client.Ping(context.Background())
if err != nil {
return err
}
if !ok {
return fmt.Errorf("connection to %s not healthy", uri)
}
return nil
}
func (s *InfluxAsyncSink) Write(m lp.CCMetric) error {
s.writeApi.WritePoint(
m.ToPoint(s.meta_as_tags),
)
s.sentMetrics++
stats.ComponentStatInt(s.name, "send_metrics", s.sentMetrics)
return nil
}
func (s *InfluxAsyncSink) Flush() error {
s.writeApi.Flush()
s.statsFlushes++
stats.ComponentStatInt(s.name, "flushes", s.statsFlushes)
return nil
}
func (s *InfluxAsyncSink) Close() {
cclog.ComponentDebug(s.name, "Closing InfluxDB connection")
s.writeApi.Flush()
s.client.Close()
}
func NewInfluxAsyncSink(name string, config json.RawMessage) (Sink, error) {
s := new(InfluxAsyncSink)
s.name = fmt.Sprintf("InfluxSink(%s)", name)
// Set default for maximum number of points sent to server in single request.
s.config.BatchSize = 0
s.influxRetryInterval = 0
//s.config.InfluxRetryInterval = "1s"
s.influxMaxRetryTime = 0
//s.config.InfluxMaxRetryTime = "168h"
s.config.InfluxMaxRetries = 0
s.config.InfluxExponentialBase = 0
s.config.FlushInterval = 0
// Default retry intervals (in seconds)
// 1 2
// 2 4
// 4 8
// 8 16
// 16 32
// 32 64
// 64 128
// 128 256
// 256 512
// 512 1024
// 1024 2048
// 2048 4096
// 4096 8192
// 8192 16384
// 16384 32768
// 32768 65536
// 65536 131072
// 131072 262144
// 262144 524288
if len(config) > 0 {
err := json.Unmarshal(config, &s.config)
if err != nil {
return nil, err
}
}
if len(s.config.Host) == 0 ||
len(s.config.Port) == 0 ||
len(s.config.Database) == 0 ||
len(s.config.Organization) == 0 ||
len(s.config.Password) == 0 {
return nil, errors.New("not all configuration variables set required by InfluxAsyncSink")
}
// Create lookup map to use meta infos as tags in the output metric
s.meta_as_tags = make(map[string]bool)
for _, k := range s.config.MetaAsTags {
s.meta_as_tags[k] = true
}
toUint := func(duration string, def uint) uint {
t, err := time.ParseDuration(duration)
if err == nil {
return uint(t.Milliseconds())
}
return def
}
s.influxRetryInterval = toUint(s.config.InfluxRetryInterval, s.influxRetryInterval)
s.influxMaxRetryTime = toUint(s.config.InfluxMaxRetryTime, s.influxMaxRetryTime)
// Connect to InfluxDB server
if err := s.connect(); err != nil {
return nil, fmt.Errorf("unable to connect: %v", err)
}
// Start background: Read from error channel
s.statsErrors = 0
s.errors = s.writeApi.Errors()
go func() {
for err := range s.errors {
s.statsErrors++
stats.ComponentStatInt(s.name, "errors", s.statsErrors)
cclog.ComponentError(s.name, err.Error())
}
}()
s.sentMetrics = 0
s.statsFlushes = 0
return s, nil
}