Skip to main content

Collector concepts

How it works

Plugins are pre-packaged sets of collector components. They simplify the installation and configuration process for the APEX AIOps Incident Management Collector. Plugins allow the collector to ingest metrics from a variety of technologies, observability tools, and infrastructures.

A plugin consists of at least one collector source and one sink, and it may also include one or more transforms. When you install an Incident Management Collector, it installs the system plugin, which will collect basic metrics about the status of the machine where the collector was installed. Once running, you can then manage plugin configuration by adding, modifying and deleting parameters and mappings. Plugins reduce the complexity of configuring sources, transforms, and sinks, and allow for management of these settings through a centralized interface so you can start viewing your data in Incident Management quickly.

The Incident Management Collector ingests observability metrics through one or more sources. A combination of transformations and a sink then normalizes the data for continued processing in Incident Management.

Figure 1. Collector Role in Metrics Ingestion
Collector Role in Metrics Ingestion


Incident Management Collector components

Metrics

Metrics are a collection of data points gathered through repeated measurements over time - a numerical operation performed across a time series.

Source

A source is a configuration defining where and how the Incident Management Collector pulls in data, or how it receives data pushed to it. Sources are specific to a particular technology and allow the collector to receive that type of data. Some examples of sources are HTTP, Mongo, Kubernetes, syslog, and Prometheus. One collector (and even one plugin) can have many sources.

Transform

A transform manipulates data after it is ingested from a source. Different transforms can perform many functions, including deduplication, data enrichment, and filtering.

Sink

A sink sends ingested log and metric data to a destination. As a core component in Vector, a sink interacts with services external to the collector while sending data to the Incident Management engine. As the terminus point for a Vector pipeline, a sink is integral to event handling, sending events to other services for additional processing, or outputting the data for visualizations and analyses.