cc-metric-collector/sinks/influxAsyncSink.go

240 lines
7.4 KiB
Go
Raw Permalink Normal View History

package sinks
import (
"context"
"crypto/tls"
"encoding/json"
"errors"
"fmt"
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-metric-collector/pkg/ccMetric"
influxdb2 "github.com/influxdata/influxdb-client-go/v2"
influxdb2Api "github.com/influxdata/influxdb-client-go/v2/api"
influxdb2ApiHttp "github.com/influxdata/influxdb-client-go/v2/api/http"
)
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"`
2022-02-10 09:43:02 +01:00
// 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"`
CustomFlushInterval string `json:"custom_flush_interval,omitempty"`
MaxRetryAttempts uint `json:"max_retry_attempts,omitempty"`
}
type InfluxAsyncSink struct {
sink
client influxdb2.Client
writeApi influxdb2Api.WriteAPI
errors <-chan error
config InfluxAsyncSinkConfig
influxRetryInterval uint
influxMaxRetryTime uint
customFlushInterval time.Duration
flushTimer *time.Timer
}
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)
2022-02-09 11:08:50 +01:00
clientOptions := influxdb2.DefaultOptions()
2022-02-10 09:43:02 +01:00
if s.config.BatchSize != 0 {
cclog.ComponentDebug(s.name, "Batch size", s.config.BatchSize)
2022-02-10 09:43:02 +01:00
clientOptions.SetBatchSize(s.config.BatchSize)
}
if s.config.FlushInterval != 0 {
cclog.ComponentDebug(s.name, "Flush interval", s.config.FlushInterval)
2022-02-10 09:43:02 +01:00
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)
}
2022-02-09 11:08:50 +01:00
clientOptions.SetTLSConfig(
&tls.Config{
InsecureSkipVerify: true,
2022-02-09 11:08:50 +01:00
},
).SetPrecision(time.Second)
2022-02-09 11:08:50 +01:00
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)
}
s.writeApi.SetWriteFailedCallback(func(batch string, err influxdb2ApiHttp.Error, retryAttempts uint) bool {
mlist := strings.Split(batch, "\n")
cclog.ComponentError(s.name, fmt.Sprintf("Failed to write batch with %d metrics %d times (max: %d): %s", len(mlist), retryAttempts, s.config.MaxRetryAttempts, err.Error()))
return retryAttempts <= s.config.MaxRetryAttempts
})
return nil
}
func (s *InfluxAsyncSink) Write(m lp.CCMetric) error {
if s.customFlushInterval != 0 && s.flushTimer == nil {
// Run a batched flush for all lines that have arrived in the defined interval
s.flushTimer = time.AfterFunc(s.customFlushInterval, func() {
if err := s.Flush(); err != nil {
cclog.ComponentError(s.name, "flush failed:", err.Error())
}
})
}
s.writeApi.WritePoint(
m.ToPoint(s.meta_as_tags),
)
return nil
}
func (s *InfluxAsyncSink) Flush() error {
cclog.ComponentDebug(s.name, "Flushing")
s.writeApi.Flush()
if s.customFlushInterval != 0 && s.flushTimer != nil {
s.flushTimer = nil
}
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)
2022-02-10 09:43:02 +01:00
// 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
s.config.CustomFlushInterval = ""
s.customFlushInterval = time.Duration(0)
s.config.MaxRetryAttempts = 1
// 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
2022-02-10 09:43:02 +01:00
if len(config) > 0 {
err := json.Unmarshal(config, &s.config)
if err != nil {
return nil, err
}
}
Merge latest development changes to main branch (#79) * Cleanup: Remove unused code * Use Golang duration parser for 'interval' and 'duration' in main config * Update handling of LIKWID headers. Download only if not already present in the system. Fixes #73 * Units with cc-units (#64) * Add option to normalize units with cc-unit * Add unit conversion to router * Add option to change unit prefix in the router * Add to MetricRouter README * Add order of operations in router to README * Use second add_tags/del_tags only if metric gets renamed * Skip disks in DiskstatCollector that have size=0 * Check readability of sensor files in TempCollector * Fix for --once option * Rename `cpu` type to `hwthread` (#69) * Rename 'cpu' type to 'hwthread' to avoid naming clashes with MetricStore and CC-Webfrontend * Collectors in parallel (#74) * Provide info to CollectorManager whether the collector can be executed in parallel with others * Split serial and parallel collectors. Read in parallel first * Update NvidiaCollector with new metrics, MIG and NvLink support (#75) * CC topology module update (#76) * Rename CPU to hardware thread, write some comments * Do renaming in other parts * Remove CpuList and SocketList function from metricCollector. Available in ccTopology * Option to use MIG UUID as subtype-id in NvidiaCollector * Option to use MIG slice name as subtype-id in NvidiaCollector * MetricRouter: Fix JSON in README * Fix for Github Action to really use the selected version * Remove Ganglia installation in runonce Action and add Go 1.18 * Fix daemon options in init script * Add separate go.mod files to use it with deprecated 1.16 * Minor updates for Makefiles * fix string comparison * AMD ROCm SMI collector (#77) * Add collector for AMD ROCm SMI metrics * Fix import path * Fix imports * Remove Board Number * store GPU index explicitly * Remove board number from description * Use http instead of ftp to download likwid * Fix serial number in rocmCollector * Improved http sink (#78) * automatic flush in NatsSink * tweak default options of HttpSink * shorter cirt. section and retries for HttpSink * fix error handling * Remove file added by mistake. * Use http instead of ftp to download likwid * Fix serial number in rocmCollector Co-authored-by: Thomas Roehl <thomas.roehl@fau.de> Co-authored-by: Holger Obermaier <40787752+ho-ob@users.noreply.github.com> Co-authored-by: Lou <lou.knauer@gmx.de>
2022-06-08 15:25:40 +02:00
if len(s.config.Port) == 0 {
return nil, errors.New("Missing port configuration required by InfluxSink")
}
if len(s.config.Database) == 0 {
return nil, errors.New("Missing database configuration required by InfluxSink")
}
if len(s.config.Organization) == 0 {
return nil, errors.New("Missing organization configuration required by InfluxSink")
}
if len(s.config.Password) == 0 {
return nil, errors.New("Missing password configuration required by InfluxSink")
}
// 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
}
2022-02-09 11:08:50 +01:00
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)
// Use a own timer for calling Flush()
if len(s.config.CustomFlushInterval) > 0 {
t, err := time.ParseDuration(s.config.CustomFlushInterval)
if err != nil {
return nil, fmt.Errorf("invalid duration in 'custom_flush_interval': %v", err)
}
s.customFlushInterval = t
}
2022-02-09 11:08:50 +01:00
// Connect to InfluxDB server
if err := s.connect(); err != nil {
return nil, fmt.Errorf("unable to connect: %v", err)
}
2022-02-09 11:08:50 +01:00
// Start background: Read from error channel
s.errors = s.writeApi.Errors()
go func() {
for err := range s.errors {
cclog.ComponentError(s.name, err.Error())
}
}()
return s, nil
}