dom
/
telegraf
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Add generic line protocol socket listener plugin for influx or graphite data format stocket input....

This commit is contained in:
Henry Hu 2016-02-03 17:17:25 +08:00
parent 9fdf75e1f7
commit 5ca182c92b
4 changed files with 557 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
plugins/inputs/all/all.go
View File

@ -35,6 +35,7 @@ import (
_ "github.com/influxdata/telegraf/plugins/inputs/rethinkdb"
_ "github.com/influxdata/telegraf/plugins/inputs/sensors"
_ "github.com/influxdata/telegraf/plugins/inputs/snmp"
_ "github.com/influxdata/telegraf/plugins/inputs/socket"
_ "github.com/influxdata/telegraf/plugins/inputs/sqlserver"
_ "github.com/influxdata/telegraf/plugins/inputs/statsd"
_ "github.com/influxdata/telegraf/plugins/inputs/system"

166
plugins/inputs/socket/README.md Normal file
View File

@ -0,0 +1,166 @@
# The Socket Service Input
## A note on UDP/IP OS Buffer sizes
If you're using UDP input and running Linux or FreeBSD, please adjust your UDP buffer
size limit, [see here for more details.](../udp/README.md#a-note-on-udpip-os-buffer-sizes)
## Configuration
Each Socket input allows the binding address, and protocol to be set.
Currently socket service input only support two line protocol formats of metric: influx and graphite.
Each input allows to use data_format to choose which format of contents will be sent to this socket service.
## Parsing Metrics
The Socket service input plugin allows measurements to be saved using the influx or graphite line protocol. By default, enabling the socket plugin will allow you to collect influx or graphite metrics and store them using the metric name as the measurement. If you send a graphite metric named `servers.localhost.cpu.loadavg.10`, it will store the full metric name as the measurement with no extracted tags.
While this default setup works, it is not the ideal way to store measurements in InfluxDB since it does not take advantage of tags. It also will not perform optimally with a large dataset sizes since queries will be forced to use regexes which is known to not scale well.
To extract tags from metrics, one or more templates must be configured to parse metrics into tags and measurements.
## Templates
Templates allow matching parts of a metric name to be used as tag keys in the stored metric. They have a similar format to graphite metric names. The values in between the separators are used as the tag keys. The location of the tag key that matches the same position as the graphite metric section is used as the value. If there is no value, the graphite portion is skipped.
The special value _measurement_ is used to define the measurement name. It can have a trailing `*` to indicate that the remainder of the metric should be used. If a _measurement_ is not specified, the full metric name is used.
### Basic Matching
`servers.localhost.cpu.loadavg.10`
* Template: `.host.resource.measurement*`
* Output: _measurement_ =`loadavg.10` _tags_ =`host=localhost resource=cpu`
### Multiple Measurement Matching
The _measurement_ can be specified multiple times in a template to provide more control over the measurement name. Multiple values
will be joined together using the _Separator_ config variable. By default, this value is `.`.
`servers.localhost.cpu.cpu0.user`
* Template: `.host.measurement.cpu.measurement`
* Output: _measurement_ = `cpu.user` _tags_ = `host=localhost cpu=cpu0`
Since '.' requires queries on measurements to be double-quoted, you may want to set this to `_` to simplify querying parsed metrics.
`servers.localhost.cpu.cpu0.user`
* Separator: `_`
* Template: `.host.measurement.cpu.measurement`
* Output: _measurement_ = `cpu_user` _tags_ = `host=localhost cpu=cpu0`
### Adding Tags
Additional tags can be added to a metric that don't exist on the received metric. You can add additional tags by specifying them after the pattern. Tags have the same format as the line protocol. Multiple tags are separated by commas.
`servers.localhost.cpu.loadavg.10`
* Template: `.host.resource.measurement* region=us-west,zone=1a`
* Output: _measurement_ = `loadavg.10` _tags_ = `host=localhost resource=cpu region=us-west zone=1a`
### Fields
A field key can be specified by using the keyword _field_. By default if no _field_ keyword is specified then the metric will be written to a field named _value_.
When using the current default engine _BZ1_, it's recommended to use a single field per value for performance reasons.
When using the _TSM1_ engine it's possible to amend measurement metrics with additional fields, e.g:
Input:
```
sensu.metric.net.server0.eth0.rx_packets 461295119435 1444234982
sensu.metric.net.server0.eth0.tx_bytes 1093086493388480 1444234982
sensu.metric.net.server0.eth0.rx_bytes 1015633926034834 1444234982
sensu.metric.net.server0.eth0.tx_errors 0 1444234982
sensu.metric.net.server0.eth0.rx_errors 0 1444234982
sensu.metric.net.server0.eth0.tx_dropped 0 1444234982
sensu.metric.net.server0.eth0.rx_dropped 0 1444234982
```
With template:
```
sensu.metric.* ..measurement.host.interface.field
```
Becomes database entry:
```
> select * from net
name: net
---------
time host interface rx_bytes rx_dropped rx_errors rx_packets tx_bytes tx_dropped tx_errors
1444234982000000000 server0 eth0 1.015633926034834e+15 0 0 4.61295119435e+11 1.09308649338848e+15 0 0
```
## Multiple Templates
One template may not match all metrics. For example, using multiple plugins with diamond will produce metrics in different formats. If you need to use multiple templates, you'll need to define a prefix filter that must match before the template can be applied.
### Filters
Filters have a similar format to templates but work more like wildcard expressions. When multiple filters would match a metric, the more specific one is chosen. Filters are configured by adding them before the template.
For example,
```
servers.localhost.cpu.loadavg.10
servers.host123.elasticsearch.cache_hits 100
servers.host456.mysql.tx_count 10
servers.host789.prod.mysql.tx_count 10
```
* `servers.*` would match all values
* `servers.*.mysql` would match `servers.host456.mysql.tx_count 10`
* `servers.localhost.*` would match `servers.localhost.cpu.loadavg`
* `servers.*.*.mysql` would match `servers.host789.prod.mysql.tx_count 10`
## Default Templates
If no template filters are defined or you want to just have one basic template, you can define a default template. This template will apply to any metric that has not already matched a filter.
```
dev.http.requests.200
prod.myapp.errors.count
dev.db.queries.count
```
* `env.app.measurement*` would create
* _measurement_=`requests.200` _tags_=`env=dev,app=http`
* _measurement_= `errors.count` _tags_=`env=prod,app=myapp`
* _measurement_=`queries.count` _tags_=`env=dev,app=db`
## Global Tags
If you need to add the same set of tags to all metrics, you can define them globally at the plugin level and not within each template description.
## Minimal Config
```
[[inputs.socket]]
bind_address = ":3003" # the bind address
protocol = "tcp" # or "udp" protocol to read via
udp_read_buffer = 8388608 # (8*1024*1024) UDP read buffer size
# Data format to consume. This can be "influx" or "graphite" (line-protocol)
# NOTE json only reads numerical measurements, strings and booleans are ignored.
data_format = "graphite"
### If matching multiple measurement files, this string will be used to join the matched values.
separator = "."
### Default tags that will be added to all metrics. These can be overridden at the template level
### or by tags extracted from metric
tags = ["region=us-east", "zone=1c"]
### Each template line requires a template pattern. It can have an optional
### filter before the template and separated by spaces. It can also have optional extra
### tags following the template. Multiple tags should be separated by commas and no spaces
### similar to the line protocol format. The can be only one default template.
### Templates support below format:
### filter + template
### filter + template + extra tag
### filter + template with field key
### default template. Ignore the first graphite component "servers"
templates = [
"*.app env.service.resource.measurement",
"stats.* .host.measurement* region=us-west,agent=sensu",
"stats2.* .host.measurement.field",
"measurement*"
]
```

70
plugins/inputs/socket/config.go Normal file
View File

@ -0,0 +1,70 @@
package socket
import "github.com/influxdata/telegraf/internal/encoding/graphite"
const (
// DefaultBindAddress is the default binding interface if none is specified.
DefaultBindAddress = ":2003"
// DefaultProtocol is the default IP protocol used by the Graphite input.
DefaultProtocol = "tcp"
// DefaultSeparator is the default join character to use when joining multiple
// measurment parts in a template.
DefaultSeparator = "."
// DefaultUDPReadBuffer is the default buffer size for the UDP listener.
// Sets the size of the operating system's receive buffer associated with
// the UDP traffic. Keep in mind that the OS must be able
// to handle the number set here or the UDP listener will error and exit.
//
// DefaultReadBuffer = 0 means to use the OS default, which is usually too
// small for high UDP performance.
//
// Increasing OS buffer limits:
// Linux: sudo sysctl -w net.core.rmem_max=<read-buffer>
// BSD/Darwin: sudo sysctl -w kern.ipc.maxsockbuf=<read-buffer>
DefaultUdpReadBuffer = 0
)
// Config represents the configuration for Graphite endpoints.
type Config struct {
BindAddress string
Protocol string
UdpReadBuffer int
graphite.InnerConfig
}
// WithDefaults takes the given config and returns a new config with any required
// default values set.
func (c *Config) WithDefaults() *Config {
d := *c
if d.BindAddress == "" {
d.BindAddress = DefaultBindAddress
}
if d.Protocol == "" {
d.Protocol = DefaultProtocol
}
if d.Separator == "" {
d.Separator = DefaultSeparator
}
if d.UdpReadBuffer == 0 {
d.UdpReadBuffer = DefaultUdpReadBuffer
}
return &d
}
// New Config instance.
func NewConfig(bindAddress, protocol string, udpReadBuffer int, separator string, tags []string, templates []string) *Config {
c := &Config{}
c.BindAddress = bindAddress
c.Protocol = protocol
c.UdpReadBuffer = udpReadBuffer
c.Separator = separator
c.Tags = tags
c.Templates = templates
return c
}

320
plugins/inputs/socket/socket.go Normal file
View File

@ -0,0 +1,320 @@
package socket
import (
"bufio"
"fmt"
"log"
"math"
"net"
"os"
"strings"
"sync"
"time"
"github.com/influxdata/telegraf"
"github.com/influxdata/telegraf/internal/encoding"
"github.com/influxdata/telegraf/internal/encoding/graphite"
"github.com/influxdata/telegraf/plugins/inputs"
)
const (
udpBufferSize = 65536
)
type tcpConnection struct {
conn net.Conn
connectTime time.Time
}
func (c *tcpConnection) Close() {
c.conn.Close()
}
// Socket represents a Socket listening service.
type Socket struct {
BindAddress string
Protocol string
UdpReadBuffer int
DataFormat string
Separator string
Tags []string
Templates []string
mu sync.Mutex
encodingParser *encoding.Parser
logger *log.Logger
config *Config
tcpConnectionsMu sync.Mutex
tcpConnections map[string]*tcpConnection
ln net.Listener
addr net.Addr
udpConn *net.UDPConn
wg sync.WaitGroup
done chan struct{}
// channel for all incoming parsed points
metricC chan telegraf.Metric
}
var sampleConfig = `
bind_address = ":2003" # the bind address
protocol = "tcp" # or "udp" protocol to read via
udp_read_buffer = 8388608 # (8*1024*1024) UDP read buffer size
# Data format to consume. This can be "influx" or "graphite" (line-protocol)
# NOTE json only reads numerical measurements, strings and booleans are ignored.
data_format = "graphite"
### If matching multiple measurement files, this string will be used to join the matched values.
separator = "."
### Default tags that will be added to all metrics. These can be overridden at the template level
### or by tags extracted from metric
tags = ["region=north-china", "zone=1c"]
### Each template line requires a template pattern. It can have an optional
### filter before the template and separated by spaces. It can also have optional extra
### tags following the template. Multiple tags should be separated by commas and no spaces
### similar to the line protocol format. The can be only one default template.
### Templates support below format:
### filter + template
### filter + template + extra tag
### filter + template with field key
### default template. Ignore the first graphite component "servers"
templates = [
"*.app env.service.resource.measurement",
"stats.* .host.measurement* region=us-west,agent=sensu",
"stats2.* .host.measurement.field",
"measurement*"
]
`
func (s *Socket) SampleConfig() string {
return sampleConfig
}
func (s *Socket) Description() string {
return "Socket read influx or graphite line-protocol metrics from tcp/udp socket"
}
// Open starts the Socket input processing data.
func (s *Socket) Start() error {
s.mu.Lock()
defer s.mu.Unlock()
c := NewConfig(s.BindAddress, s.Protocol, s.UdpReadBuffer, s.Separator, s.Tags, s.Templates)
c.WithDefaults()
if err := c.Validate(); err != nil {
return fmt.Errorf("Socket input configuration is error! ", err.Error())
}
s.config = c
graphiteParser, err := graphite.NewParserWithOptions(graphite.Options{
Templates: s.config.Templates,
DefaultTags: s.config.DefaultTags(),
Separator: s.config.Separator})
if err != nil {
return fmt.Errorf("Socket input parser config is error: %s ", err.Error())
}
s.encodingParser = encoding.NewParser(graphiteParser)
s.tcpConnections = make(map[string]*tcpConnection)
s.done = make(chan struct{})
s.metricC = make(chan telegraf.Metric, 50000)
if strings.ToLower(s.config.Protocol) == "tcp" {
s.addr, err = s.openTCPServer()
} else if strings.ToLower(s.config.Protocol) == "udp" {
s.addr, err = s.openUDPServer()
} else {
return fmt.Errorf("unrecognized Socket input protocol %s", s.config.Protocol)
}
if err != nil {
return err
}
s.logger.Printf("Socket Plugin Listening on %s: %s", strings.ToUpper(s.config.Protocol), s.config.BindAddress)
return nil
}
func (s *Socket) closeAllConnections() {
s.tcpConnectionsMu.Lock()
defer s.tcpConnectionsMu.Unlock()
for _, c := range s.tcpConnections {
c.Close()
}
}
// Close stops all data processing on the Socket input.
func (s *Socket) Stop() {
s.mu.Lock()
defer s.mu.Unlock()
s.closeAllConnections()
if s.ln != nil {
s.ln.Close()
}
if s.udpConn != nil {
s.udpConn.Close()
}
close(s.done)
s.wg.Wait()
s.done = nil
}
// openTCPServer opens the Socket input in TCP mode and starts processing data.
func (s *Socket) openTCPServer() (net.Addr, error) {
ln, err := net.Listen("tcp", s.config.BindAddress)
if err != nil {
return nil, err
}
s.ln = ln
s.wg.Add(1)
go func() {
defer s.wg.Done()
for {
conn, err := s.ln.Accept()
if opErr, ok := err.(*net.OpError); ok && !opErr.Temporary() {
s.logger.Println("Socket TCP listener closed")
return
}
if err != nil {
s.logger.Println("error accepting TCP connection", err.Error())
continue
}
s.wg.Add(1)
go s.handleTCPConnection(conn)
}
}()
return ln.Addr(), nil
}
// handleTCPConnection services an individual TCP connection for the Socket input.
func (s *Socket) handleTCPConnection(conn net.Conn) {
defer s.wg.Done()
defer conn.Close()
defer s.untrackConnection(conn)
s.trackConnection(conn)
reader := bufio.NewReader(conn)
for {
// Read up to the next newline.
buf, err := reader.ReadBytes('\n')
if err != nil {
return
}
// Trim the buffer, even though there should be no padding
// line := strings.TrimSpace(string(buf))
s.handleLines(buf)
}
}
func (s *Socket) trackConnection(c net.Conn) {
s.tcpConnectionsMu.Lock()
defer s.tcpConnectionsMu.Unlock()
s.tcpConnections[c.RemoteAddr().String()] = &tcpConnection{
conn: c,
connectTime: time.Now().UTC(),
}
}
func (s *Socket) untrackConnection(c net.Conn) {
s.tcpConnectionsMu.Lock()
defer s.tcpConnectionsMu.Unlock()
delete(s.tcpConnections, c.RemoteAddr().String())
}
// openUDPServer opens the Socket input in UDP mode and starts processing incoming data.
func (s *Socket) openUDPServer() (net.Addr, error) {
addr, err := net.ResolveUDPAddr("udp", s.config.BindAddress)
if err != nil {
return nil, err
}
s.udpConn, err = net.ListenUDP("udp", addr)
if err != nil {
return nil, err
}
if s.config.UdpReadBuffer != 0 {
err = s.udpConn.SetReadBuffer(s.config.UdpReadBuffer)
if err != nil {
return nil, fmt.Errorf("unable to set UDP read buffer to %d: %s",
s.config.UdpReadBuffer, err)
}
}
buf := make([]byte, udpBufferSize)
s.wg.Add(1)
go func() {
defer s.wg.Done()
for {
n, _, err := s.udpConn.ReadFromUDP(buf)
if err != nil {
s.udpConn.Close()
return
}
s.handleLines(buf[:n])
}
}()
return s.udpConn.LocalAddr(), nil
}
func (s *Socket) handleLines(buf []byte) {
if buf == nil || len(buf) < 1 {
return
}
// Parse it.
metrics, err := s.encodingParser.ParseSocketLines(s.DataFormat, buf)
if err != nil {
switch err := err.(type) {
case *graphite.UnsupposedValueError:
// Socket ignores NaN values with no error.
if math.IsNaN(err.Value) {
return
}
}
s.logger.Printf("unable to parse lines: %s: %s", buf, err)
return
}
for _, metric := range metrics {
s.metricC <- metric
}
}
func (s *Socket) Gather(acc telegraf.Accumulator) error {
s.mu.Lock()
defer s.mu.Unlock()
npoints := len(s.metricC)
for i := 0; i < npoints; i++ {
metric := <-s.metricC
acc.AddFields(metric.Name(), metric.Fields(), metric.Tags(), metric.Time())
}
return nil
}
func init() {
inputs.Add("socket", func() telegraf.Input {
return &Socket{logger: log.New(os.Stderr, "[socket] ", log.LstdFlags)}
})
}