AI4ME: stats4me

A telemetry stack for use in the research & development of OBM services.

---

Overview

• Collection: This stack uses Telegraf to collect metrics from the local system via "inputs" and export them to the data store.

• Storage: InfluxDB is a time-series database that simply stores timestamped (and labeled) data that can be queried using the flux query language.

• Visualization: For pretty graphs, Grafana can house web-based dashboards that contain graphs etc that visualize data. It can be set to collect data from InfluxDB via flux queries.

Usage

This is a docker compose stack that will download run all the required tools using docker and docker compose.

• Downloading: docker compose pull
• Running: docker compose up -d
• Logs: docker compose logs -f
• Stop: docker compose down

Accessing these services is done via a browser on the localhost. InfluxDB's UI can be accessed at port 8086 by default and Grafana at port 3000 (again, by default).

Configuration

Each service has a configuration, and each can be complex to configure if used in complex ways. A basic configuration setup for each is generated by this repository.

InfluxDB

To get started with InfluxDB, start with providing an initial config. In this repo, this can be done via env vars found in the .env file.

• INFLUXDB_USERNAME: Is the username for accessing the default InfluxDB account.
• INFLUXDB_PASSWORD: Is the password for accessing the default InfluxDB account.
• INFLUXDB_BUCKET: Is the name of the initial bucket that can be used to store data in.
• INFLUXDB_ORG: Is the name of the initial organization in the InfluxDB.
• INFLUXDB_TOKEN: The admin API token used to get and push data to the database.

These values are also used by telegraf in the initial config so that it can be autoconfigured to have an influx output to the initial bucket.

Telegraf

Telegraf is configured by a toml file that can be found at telegraf/telegraf.conf. Initially, this is setup with a few basic input plugins (such as docker) and it outputs data to the influxdb instance at the initial bucket. For more about telegraf plugins and configuration, see here.

So Telegraf can collect data from the docker engine, it needs to be running in the docker group. For this, set the DOCKER_GID in the .env file to the value found for the docker group ID. This can be found by running id in the command line (provided your current user is in the docker group too).

Grafana

Grafana has very little configuration done by this repository... The default credentials are unchanged (typically: admin, admin). For more on Grafana config, see here.