Common Concepts of OpenTSDB Application Development

Basic Concepts

data point: A time series data point consists of a metric, a timestamp, a value, and a set of tags. The data point indicates the value of a metric at a specific time point.
metric: Metrics include CPU usage, memory, and I/Os in system monitoring.
timestamp: A UNIX timestamp (seconds or milliseconds since Epoch), that is, the time when the value is generated.
value: The value of a metric is a JSON formatted event or a histogram/digest.
tag: A tag is a key-value pair consisting of Tagk and Tagv. It describes the time series the point belongs to.

Tags allow you to separate similar data points from different sources or related entities, so you can easily graph them individually or in groups. One common use case for tags consists in annotating a data point with the name of the machine that produced it as well as name of the cluster or pool the machine belongs to. This allows you to easily make dashboards that show the state of your service on a per-server basis as well as dashboards that show an aggregated state across logical pools of servers.

Introduction to an OpenTSDB System Table

OpenTSDB stores time series data based on HBase. After OpenTSDB is enabled in a cluster, the system will create four HBase tables in the cluster. Table 1 describes the OpenTSDB system tables.

Do not modify the four HBase tables manually, because this may cause unavailable OpenTSDB.

**Table 1** OpenTSDB system table
Table Name	Description
tsdb	It stores data points. All OpenTSDB data is stored in this table.
tsdb-meta	It stores time series indexes and metadata.
tsdb-tree	It stores metric structure information.
tsdb-uid	It stores unique identifier (UID) mappings. Each metric in a data point is mapped to a UID, and each tag in a data point is mapped to a UID. At the same time, each UID is reversely mapped to the metric or tag. These mappings are stored in this table.