Telegraf is able to parse the following input data formats into metrics:
- InfluxDB Line Protocol
- JSON
- Graphite
- Value, ie: 45 or "booyah"
- Nagios (exec input only)
- Collectd
- Dropwizard
Telegraf metrics, like InfluxDB points, are a combination of four basic parts:
- Measurement Name
- Tags
- Fields
- Timestamp
These four parts are easily defined when using InfluxDB line-protocol as a data format. But there are other data formats that users may want to use which require more advanced configuration to create usable Telegraf metrics.
Plugins such as exec
and kafka_consumer
parse textual data. Up until now,
these plugins were statically configured to parse just a single
data format. exec
mostly only supported parsing JSON, and kafka_consumer
only
supported data in InfluxDB line-protocol.
But now we are normalizing the parsing of various data formats across all
plugins that can support it. You will be able to identify a plugin that supports
different data formats by the presence of a data_format
config option, for
example, in the exec plugin:
[[inputs.exec]]
## Commands array
commands = ["/tmp/test.sh", "/usr/bin/mycollector --foo=bar"]
## measurement name suffix (for separating different commands)
name_suffix = "_mycollector"
## Data format to consume.
## Each data format has its own unique set of configuration options, read
## more about them here:
## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
data_format = "json"
## Additional configuration options go here
Each data_format has an additional set of configuration options available, which I'll go over below.
There are no additional configuration options for InfluxDB line-protocol. The metrics are parsed directly into Telegraf metrics.
[[inputs.exec]]
## Commands array
commands = ["/tmp/test.sh", "/usr/bin/mycollector --foo=bar"]
## measurement name suffix (for separating different commands)
name_suffix = "_mycollector"
## Data format to consume.
## Each data format has its own unique set of configuration options, read
## more about them here:
## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
data_format = "influx"
The JSON data format flattens JSON into metric fields. NOTE: Only numerical values are converted to fields, and they are converted into a float. strings are ignored unless specified as a tag_key (see below).
So for example, this JSON:
{
"a": 5,
"b": {
"c": 6
},
"ignored": "I'm a string"
}
Would get translated into fields of a measurement:
myjsonmetric a=5,b_c=6
The measurement name is usually the name of the plugin,
but can be overridden using the name_override
config option.
The JSON data format supports specifying "tag keys". If specified, keys will be searched for in the root-level of the JSON blob. If the key(s) exist, they will be applied as tags to the Telegraf metrics.
For example, if you had this configuration:
[[inputs.exec]]
## Commands array
commands = ["/tmp/test.sh", "/usr/bin/mycollector --foo=bar"]
## measurement name suffix (for separating different commands)
name_suffix = "_mycollector"
## Data format to consume.
## Each data format has its own unique set of configuration options, read
## more about them here:
## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
data_format = "json"
## List of tag names to extract from top-level of JSON server response
tag_keys = [
"my_tag_1",
"my_tag_2"
]
with this JSON output from a command:
{
"a": 5,
"b": {
"c": 6
},
"my_tag_1": "foo"
}
Your Telegraf metrics would get tagged with "my_tag_1"
exec_mycollector,my_tag_1=foo a=5,b_c=6
If the JSON data is an array, then each element of the array is parsed with the configured settings. Each resulting metric will be output with the same timestamp.
For example, if the following configuration:
[[inputs.exec]]
## Commands array
commands = ["/usr/bin/mycollector --foo=bar"]
## measurement name suffix (for separating different commands)
name_suffix = "_mycollector"
## Data format to consume.
## Each data format has its own unique set of configuration options, read
## more about them here:
## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
data_format = "json"
## List of tag names to extract from top-level of JSON server response
tag_keys = [
"my_tag_1",
"my_tag_2"
]
with this JSON output from a command:
[
{
"a": 5,
"b": {
"c": 6
},
"my_tag_1": "foo",
"my_tag_2": "baz"
},
{
"a": 7,
"b": {
"c": 8
},
"my_tag_1": "bar",
"my_tag_2": "baz"
}
]
Your Telegraf metrics would get tagged with "my_tag_1" and "my_tag_2"
exec_mycollector,my_tag_1=foo,my_tag_2=baz a=5,b_c=6
exec_mycollector,my_tag_1=bar,my_tag_2=baz a=7,b_c=8
The "value" data format translates single values into Telegraf metrics. This is done by assigning a measurement name and setting a single field ("value") as the parsed metric.
You must tell Telegraf what type of metric to collect by using the
data_type
configuration option. Available options are:
- integer
- float or long
- string
- boolean
Note: It is also recommended that you set name_override
to a measurement
name that makes sense for your metric, otherwise it will just be set to the
name of the plugin.
[[inputs.exec]]
## Commands array
commands = ["cat /proc/sys/kernel/random/entropy_avail"]
## override the default metric name of "exec"
name_override = "entropy_available"
## Data format to consume.
## Each data format has its own unique set of configuration options, read
## more about them here:
## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
data_format = "value"
data_type = "integer" # required
The Graphite data format translates graphite dot buckets directly into telegraf measurement names, with a single value field, and without any tags. By default, the separator is left as ".", but this can be changed using the "separator" argument. For more advanced options, Telegraf supports specifying "templates" to translate graphite buckets into Telegraf metrics.
Templates are of the form:
"host.mytag.mytag.measurement.measurement.field*"
Where the following keywords exist:
measurement
: specifies that this section of the graphite bucket corresponds to the measurement name. This can be specified multiple times.field
: specifies that this section of the graphite bucket corresponds to the field name. This can be specified multiple times.measurement*
: specifies that all remaining elements of the graphite bucket correspond to the measurement name.field*
: specifies that all remaining elements of the graphite bucket correspond to the field name.
Any part of the template that is not a keyword is treated as a tag key. This can also be specified multiple times.
NOTE: field*
cannot be used in conjunction with measurement*
!
The most basic template is to specify a single transformation to apply to all incoming metrics. So the following template:
templates = [
"region.region.measurement*"
]
would result in the following Graphite -> Telegraf transformation.
us.west.cpu.load 100
=> cpu.load,region=us.west value=100
Multiple templates can also be specified, but these should be differentiated using filters (see below for more details)
templates = [
"*.*.* region.region.measurement", # <- all 3-part measurements will match this one.
"*.*.*.* region.region.host.measurement", # <- all 4-part measurements will match this one.
]
The field keyword tells Telegraf to give the metric that field name. So the following template:
separator = "_"
templates = [
"measurement.measurement.field.field.region"
]
would result in the following Graphite -> Telegraf transformation.
cpu.usage.idle.percent.eu-east 100
=> cpu_usage,region=eu-east idle_percent=100
The field key can also be derived from all remaining elements of the graphite
bucket by specifying field*
:
separator = "_"
templates = [
"measurement.measurement.region.field*"
]
which would result in the following Graphite -> Telegraf transformation.
cpu.usage.eu-east.idle.percentage 100
=> cpu_usage,region=eu-east idle_percentage=100
Users can also filter the template(s) to use based on the name of the bucket, using glob matching, like so:
templates = [
"cpu.* measurement.measurement.region",
"mem.* measurement.measurement.host"
]
which would result in the following transformation:
cpu.load.eu-east 100
=> cpu_load,region=eu-east value=100
mem.cached.localhost 256
=> mem_cached,host=localhost value=256
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.
templates = [
"measurement.measurement.field.region datacenter=1a"
]
would result in the following Graphite -> Telegraf transformation.
cpu.usage.idle.eu-east 100
=> cpu_usage,region=eu-east,datacenter=1a idle=100
There are many more options available, More details can be found here
[[inputs.exec]]
## Commands array
commands = ["/tmp/test.sh", "/usr/bin/mycollector --foo=bar"]
## measurement name suffix (for separating different commands)
name_suffix = "_mycollector"
## Data format to consume.
## Each data format has its own unique set of configuration options, read
## more about them here:
## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
data_format = "graphite"
## This string will be used to join the matched values.
separator = "_"
## 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. There can be only one default template.
## Templates support below format:
## 1. filter + template
## 2. filter + template + extra tag(s)
## 3. filter + template with field key
## 4. default template
templates = [
"*.app env.service.resource.measurement",
"stats.* .host.measurement* region=eu-east,agent=sensu",
"stats2.* .host.measurement.field",
"measurement*"
]
There are no additional configuration options for Nagios line-protocol. The metrics are parsed directly into Telegraf metrics.
Note: Nagios Input Data Formats is only supported in exec
input plugin.
[[inputs.exec]]
## Commands array
commands = ["/usr/lib/nagios/plugins/check_load -w 5,6,7 -c 7,8,9"]
## measurement name suffix (for separating different commands)
name_suffix = "_mycollector"
## Data format to consume.
## Each data format has its own unique set of configuration options, read
## more about them here:
## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
data_format = "nagios"
The collectd format parses the collectd binary network protocol. Tags are created for host, instance, type, and type instance. All collectd values are added as float64 fields.
For more information about the binary network protocol see here.
You can control the cryptographic settings with parser options. Create an
authentication file and set collectd_auth_file
to the path of the file, then
set the desired security level in collectd_security_level
.
Additional information including client setup can be found here.
You can also change the path to the typesdb or add additional typesdb using
collectd_typesdb
.
[[inputs.socket_listener]]
service_address = "udp://127.0.0.1:25826"
name_prefix = "collectd_"
## Data format to consume.
## Each data format has its own unique set of configuration options, read
## more about them here:
## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
data_format = "collectd"
## Authentication file for cryptographic security levels
collectd_auth_file = "/etc/collectd/auth_file"
## One of none (default), sign, or encrypt
collectd_security_level = "encrypt"
## Path of to TypesDB specifications
collectd_typesdb = ["/usr/share/collectd/types.db"]
The dropwizard format can parse the JSON representation of a single dropwizard metric registry. By default, tags are parsed from metric names as if they were actual influxdb line protocol keys (measurement<,tag_set>
) which can be overriden by defining custom measurement & tag templates. All field value types are supported, string
, number
and boolean
.
A typical JSON of a dropwizard metric registry:
{
"version": "3.0.0",
"counters" : {
"measurement,tag1=green" : {
"count" : 1
}
},
"meters" : {
"measurement" : {
"count" : 1,
"m15_rate" : 1.0,
"m1_rate" : 1.0,
"m5_rate" : 1.0,
"mean_rate" : 1.0,
"units" : "events/second"
}
},
"gauges" : {
"measurement" : {
"value" : 1
}
},
"histograms" : {
"measurement" : {
"count" : 1,
"max" : 1.0,
"mean" : 1.0,
"min" : 1.0,
"p50" : 1.0,
"p75" : 1.0,
"p95" : 1.0,
"p98" : 1.0,
"p99" : 1.0,
"p999" : 1.0,
"stddev" : 1.0
}
},
"timers" : {
"measurement" : {
"count" : 1,
"max" : 1.0,
"mean" : 1.0,
"min" : 1.0,
"p50" : 1.0,
"p75" : 1.0,
"p95" : 1.0,
"p98" : 1.0,
"p99" : 1.0,
"p999" : 1.0,
"stddev" : 1.0,
"m15_rate" : 1.0,
"m1_rate" : 1.0,
"m5_rate" : 1.0,
"mean_rate" : 1.0,
"duration_units" : "seconds",
"rate_units" : "calls/second"
}
}
}
Would get translated into 4 different measurements:
measurement,metric_type=counter,tag1=green count=1
measurement,metric_type=meter count=1,m15_rate=1.0,m1_rate=1.0,m5_rate=1.0,mean_rate=1.0
measurement,metric_type=gauge value=1
measurement,metric_type=histogram count=1,max=1.0,mean=1.0,min=1.0,p50=1.0,p75=1.0,p95=1.0,p98=1.0,p99=1.0,p999=1.0
measurement,metric_type=timer count=1,max=1.0,mean=1.0,min=1.0,p50=1.0,p75=1.0,p95=1.0,p98=1.0,p99=1.0,p999=1.0,stddev=1.0,m15_rate=1.0,m1_rate=1.0,m5_rate=1.0,mean_rate=1.0
You may also parse a dropwizard registry from any JSON document which contains a dropwizard registry in some inner field. Eg. to parse the following JSON document:
{
"time" : "2017-02-22T14:33:03.662+02:00",
"tags" : {
"tag1" : "green",
"tag2" : "yellow"
},
"metrics" : {
"counters" : {
"measurement" : {
"count" : 1
}
},
"meters" : {},
"gauges" : {},
"histograms" : {},
"timers" : {}
}
}
and translate it into:
measurement,metric_type=counter,tag1=green,tag2=yellow count=1 1487766783662000000
you simply need to use the following additional configuration properties:
dropwizard_metric_registry_path = "metrics"
dropwizard_time_path = "time"
dropwizard_time_format = "2006-01-02T15:04:05Z07:00"
dropwizard_tags_path = "tags"
## tag paths per tag are supported too, eg.
#[inputs.yourinput.dropwizard_tag_paths]
# tag1 = "tags.tag1"
# tag2 = "tags.tag2"
For more information about the dropwizard json format see here.
[[inputs.exec]]
## Commands array
commands = ["curl http://localhost:8080/sys/metrics"]
timeout = "5s"
## Data format to consume.
## Each data format has its own unique set of configuration options, read
## more about them here:
## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
data_format = "dropwizard"
## Used by the templating engine to join matched values when cardinality is > 1
separator = "_"
## 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. There can be only one default template.
## Templates support below format:
## 1. filter + template
## 2. filter + template + extra tag(s)
## 3. filter + template with field key
## 4. default template
## By providing an empty template array, templating is disabled and measurements are parsed as influxdb line protocol keys (measurement<,tag_set>)
templates = []
## You may use an appropriate [gjson path](https://github.com/tidwall/gjson#path-syntax)
## to locate the metric registry within the JSON document
# dropwizard_metric_registry_path = "metrics"
## You may use an appropriate [gjson path](https://github.com/tidwall/gjson#path-syntax)
## to locate the default time of the measurements within the JSON document
# dropwizard_time_path = "time"
# dropwizard_time_format = "2006-01-02T15:04:05Z07:00"
## You may use an appropriate [gjson path](https://github.com/tidwall/gjson#path-syntax)
## to locate the tags map within the JSON document
# dropwizard_tags_path = "tags"
## You may even use tag paths per tag
# [inputs.exec.dropwizard_tag_paths]
# tag1 = "tags.tag1"
# tag2 = "tags.tag2"