IPMI Sensor and Apache Druid Integration

Input and output integration overview

The IPMI Sensor Plugin facilitates the collection of server health metrics directly from hardware via the IPMI protocol, querying sensor data from either local or remote systems.

This plugin allows Telegraf to send JSON-formatted metrics to Apache Druid over HTTP, enabling real-time ingestion for analytical queries on high-volume time-series data.

Integration details

IPMI Sensor

The IPMI Sensor plugin is designed to gather bare metal metrics via the command line utility ipmitool, which interfaces with the Intelligent Platform Management Interface (IPMI). This protocol provides management and monitoring capabilities for hardware components in server systems, allowing for the retrieval of critical system health metrics such as temperature, fan speeds, and power supply status from both local and remote servers. When configured without specified servers, the plugin defaults to querying the local machine’s sensor statistics using the ipmitool sdr command. In scenarios covering remote hosts, authentication is supported through username and password using the command format ipmitool -I lan -H SERVER -U USERID -P PASSW0RD sdr. This flexibility allows users to monitor systems effectively across various environments. The plugin also supports multiple sensor types, including chassis power status and DCMI power readings, catering to administrators needing real-time insight into server operations.

Apache Druid

This configuration uses Telegraf’s HTTP output plugin with json data format to send metrics directly to Apache Druid, a real-time analytics database designed for fast, ad hoc queries on high-ingest time-series data. Druid supports ingestion via HTTP POST to various components like the Tranquility service or native ingestion endpoints. The JSON format is ideal for structuring Telegraf metrics into event-style records for Druid’s columnar and time-partitioned storage engine. Druid excels at powering interactive dashboards and exploratory queries across massive datasets, making it an excellent choice for real-time observability and monitoring analytics when integrated with Telegraf.

Configuration

IPMI Sensor

[[inputs.ipmi_sensor]]
  ## Specify the path to the ipmitool executable
  # path = "/usr/bin/ipmitool"

  ## Use sudo
  ## Setting 'use_sudo' to true will make use of sudo to run ipmitool.
  ## Sudo must be configured to allow the telegraf user to run ipmitool
  ## without a password.
  # use_sudo = false

  ## Servers
  ## Specify one or more servers via a url. If no servers are specified, local
  ## machine sensor stats will be queried. Uses the format:
  ##  [username[:password]@][protocol[(address)]]
  ##  e.g. root:passwd@lan(127.0.0.1)
  # servers = ["USERID:PASSW0RD@lan(192.168.1.1)"]

  ## Session privilege level
  ## Choose from: CALLBACK, USER, OPERATOR, ADMINISTRATOR
  # privilege = "ADMINISTRATOR"

  ## Timeout
  ## Timeout for the ipmitool command to complete.
  # timeout = "20s"

  ## Metric schema version
  ## See the plugin readme for more information on schema versioning.
  # metric_version = 1

  ## Sensors to collect
  ## Choose from:
  ##   * sdr: default, collects sensor data records
  ##   * chassis_power_status: collects the power status of the chassis
  ##   * dcmi_power_reading: collects the power readings from the Data Center Management Interface
  # sensors = ["sdr"]

  ## Hex key
  ## Optionally provide the hex key for the IMPI connection.
  # hex_key = ""

  ## Cache
  ## If ipmitool should use a cache
  ## Using a cache can speed up collection times depending on your device.
  # use_cache = false

  ## Path to the ipmitools cache file (defaults to OS temp dir)
  ## The provided path must exist and must be writable
  # cache_path = ""

Apache Druid

[[outputs.http]]
  ## Druid ingestion endpoint (e.g., Tranquility, HTTP Ingest, or Kafka REST Proxy)
  url = "http://druid-ingest.example.com/v1/post"

  ## Use POST method to send events
  method = "POST"

  ## Data format for Druid ingestion (expects JSON format)
  data_format = "json"

  ## Optional headers (may vary depending on Druid setup)
  # [outputs.http.headers]
  #   Content-Type = "application/json"
  #   Authorization = "Bearer YOUR_API_TOKEN"

  ## Optional timeout and TLS settings
  timeout = "10s"
  # tls_ca = "/path/to/ca.pem"
  # tls_cert = "/path/to/cert.pem"
  # tls_key = "/path/to/key.pem"
  # insecure_skip_verify = false

Input and output integration examples

IPMI Sensor

1. Centralized Monitoring Dashboard: Utilize the IPMI Sensor plugin to gather metrics from multiple servers and compile them into a centralized monitoring dashboard. This enables real-time visibility into server health across data centers. Administrators can track metrics like temperature and power usage, helping them make data-driven decisions about resource allocation, potential failures, and maintenance schedules.

2. Automated Power Alerts: Incorporate the plugin into an alerting system that monitors chassis power status and triggers alerts when anomalies are detected. For instance, if the power status indicates a failure or if watt values exceed expected thresholds, automated notifications can be sent to operations teams, ensuring prompt attention to hardware issues.

3. Energy Consumption Analysis: Leverage the DCMI power readings collected via the plugin to analyze energy consumption patterns of hardware over time. By integrating these readings with analytics platforms, organizations can identify opportunities to reduce power usage, optimize efficiency, and potentially decrease operational costs in large server farms or cloud infrastructures.

4. Health Check Automation: Schedule regular health checks by using the IPMI Sensor Plugin to collect data from a fleet of servers. This data can be logged and compared against historical performance metrics to identify trends, outliers, or signs of impending hardware failure, allowing IT teams to take proactive measures and reduce downtime.

Apache Druid

1. Real-Time Application Monitoring Dashboard: Use Telegraf to collect metrics from application servers and send them to Druid for immediate analysis and visualization in dashboards. Druid’s low-latency querying allows users to interactively explore system behavior in near real-time.

2. Security Event Aggregation: Aggregate and forward security-related metrics such as failed logins, port scans, or process anomalies to Druid. Analysts can build dashboards to monitor threat patterns and investigate incidents with millisecond-level granularity.

3. IoT Device Analytics: Collect telemetry from edge devices via Telegraf and send it to Druid for fast, scalable processing. Druid’s time-partitioned storage and roll-up capabilities are ideal for handling billions of small JSON events from sensors or gateways.

4. Web Traffic Behavior Exploration: Use Telegraf to capture web server metrics (e.g., requests per second, latency, error rates) and forward them to Druid. This enables teams to drill down into user behavior by region, device, or request type with subsecond query performance.

Feedback

Thank you for being part of our community! If you have any general feedback or found any bugs on these pages, we welcome and encourage your input. Please submit your feedback in the InfluxDB community Slack.