drpsychick/telegraf repository overview

⁠Docker image: drpsychick/telegraf⁠

based on telegraf:alpine docker image

⁠Purpose

• make it fully configurable through environment variables
• use one image to run them all
• run stateless, environment configured containers (see https://12factor.net/⁠)

⁠Usage

⁠UPDATE 2021-04-28

BC Breaking Change in the environments:

The image is now using a small go utility (toml_update) to read, modify and write a valid toml configuration file.

# before
TEL_AGENT_HOSTNAME=hostname = "myhostname"
TEL_OUTPUTS_INFLUXDB_URLS=urls = ["http://yourinfluxhost:8086"]
TEL_INPUTS_CPU_FLAGS = "percpu = true\ntotalcpu = true"
# after
TEL_AGENT_NAMEDOESNOTMATTER=agent.hostname=myhostname
TEL_REALLYDOESNOTMATTER=outputs.influxdb.urls=["https://yourinfluxhost:8086"]
TEL_CPUFLAG1=inputs.cpu.percpu=true
TEL_CPUFLAG2=inputs.cpu.totalcpu=true

Try it in 3 steps

⁠1 create your own telegraf.env

docker run --rm -it drpsychick/telegraf:latest cat /default.env > telegraf.env

⁠2 configure it

Edit at least your hostname and output (influxdb or sth. else) in telegraf.env:

TLG_AGENT_HOSTNAME=agent.hostname="myhostname"
TLG_INFLUXDB_URL=outputs.influxdb.urls=["http://yourinfluxhost:8086"]

You can add as many variables as you want for more inputs and their configuration, there are only a few rules:

1. the variable must start with a known prefix (the CONF_PREFIX variable in default.env)
2. the value must be a single row - due to Docker environment variables restrictions
3. use one variable for each setting

For more examples see default.env

TLG_INPUTS_CPU_PERCPU=inputs.cpu.percpu=true
TLG_INPUTS_CPU_TOTAL=inputs.cpu.totalcpu=true
TLG_INPUTS_CPU_TIME=inputs.cpu.collect_cpu_time=false
TLG_INPUTS_CPU_ACTIVE=inputs.cpu.report_active=false

The toml_update tool will take the configured prefixes in order and update the configuration file from all matching variables one-by-one.

⁠3 test and run it

Run in a separate teminal

docker run --rm -it --cap-add NET_ADMIN --env-file telegraf.env --name telegraf-1 drpsychick/telegraf:latest telegraf --test
docker run --rm -it --cap-add NET_ADMIN --env-file telegraf.env --name telegraf-1 drpsychick/telegraf:latest

Check your influxdb for new input

⁠Configure it to your needs

You can use any TLG_ variable in your telegraf.env. They will be added to the config during container startup.

⁠Example

TLG_INPUTS_DISK_FLAGS=inputs.disk.ignore_fs=["tmpfs", "devtmpfs", "devfs"]

Beware:

Docker only support simple variables. No ", no ' and especially no newlines in variables.