rrdtool

rrdtool [options]

create
    Creates a new RRD database.

update
    Updates an RRD database with new data.

fetch
    Fetches data from an RRD database.

graph
    Generates a graph from RRD data.

info
    Displays information about an RRD database.

tune
    Modify the header of an existing RRD file

dump
    Dumps the contents of an RRD database in XML format.

restore
    Restores data from an XML dump into an RRD database.

resize
    Change the number of rows in an RRA.

first
    Determines the timestamp of the first entry in an RRD.

last
    Determines the timestamp of the last entry in an RRD.

lastupdate
    Returns the time of the last update of a RRD file.

xport
    Exports data in XML or JSON.

rrdtool is a powerful suite of tools for collecting, storing, and graphing time-series data. It uses a Round Robin Database (RRD) to store data in a fixed-size database, automatically pruning older data as new data is added, ensuring a consistent storage footprint. This makes it ideal for monitoring system performance metrics, network traffic, sensor data, and any other numerical data that changes over time. rrdtool provides commands for creating databases, updating data, fetching data, and generating graphs in various formats like PNG, SVG, and PDF. The graphing capabilities are highly customizable, allowing users to create detailed visualizations of their time-series data with labels, legends, and annotations. RRDtool is used extensively in system administration, network monitoring, and scientific data analysis.
The core principle behind RRDtool is to provide efficient storage and retrieval of time-series data while maintaining a fixed database size regardless of the monitoring duration. This enables long-term trend analysis without unbounded storage requirements. The data is stored using consolidation functions, averaging or summarizing data points over time, to reduce storage overhead for older data.

Data loss can occur if the RRD is not properly configured. Understanding consolidation functions is crucial for accurate long-term data analysis. Be careful about update intervals to ensure your data matches your expectations.

Data Sources define what type of data will be stored in the RRD. Common types include GAUGE, COUNTER, DERIVE, and ABSOLUTE. Choosing the correct DS type is crucial for accurate data representation.

RRAs define how data is consolidated over time. Different RRAs can store data at different resolutions (e.g., every minute, every hour, every day). This allows you to store detailed data for recent periods and less detailed data for older periods.

rrdtool graph offers a huge set of options to customize the visual presentation. Common options are TITLE, TEXTALIGN, X-AXIS, Y-AXIS and many commands to create complex and beautiful graphs.

rrdtool was created by Tobias Oetiker. The initial version was written to visualize network traffic data collected using MRTG (Multi Router Traffic Grapher). Over time, rrdtool evolved into a standalone tool capable of handling various time-series data sources. Its widespread adoption led to its integration in numerous monitoring systems and data analysis platforms. The development of rrdtool focused on efficiency and scalability, enabling it to manage large volumes of data with minimal resource consumption.

sar(1), vmstat(8), top(1)