Opsbox Plugin System Overview

Welcome to Opsbox, where our plugin-based architecture makes cloud infrastructure management seamless and efficient! Let’s dive into how Opsbox uses its flexible plugin system to keep everything composable and extendable.

Types of Plugins

At the core of Opsbox are handlers, plugins that specify and handle other types of plugins. They implement functions that can handle many plugin types and structures.

Right now we have 2 handlers built, the general and rego handlers.

Let's go over these handlers and what they can be used for.

General Handler

These plugins use the general handler, which handles code not specifically related to rego checks and general-use plugins that don't need to define their own hookspecs.

1. Providers

Providers act as data sources (e.g., EC2, S3, etc.). They gather information from external sources and forward them to plugins that use them.

2. Inputs

Inputs can collect data from a plugin and process them or produce their own data, but they are always the start of a pipeline. They provide formatted results that other plugins can use.

3. Assistants

Assistants process, transform, and enhance the results from inputs. They act as intermediaries, refining the data for better analysis and decision-making.

4. Outputs

Outputs handle the final formatted results, directing them to specified destinations (e.g., Jira, text files, etc.).

Rego Handler

Allows for connection and processing of rego-formatted inputs called rego checks into the pipeline. Rego-based plugins use the Open Policy Agent to execute and gather details from associated rego code.

Check

A rego check collects data from a provider and executes given rego code on that data using a trusted OPA server.

By combining these plugin types, you can create a seamless pipeline for automatically generating and managing cloud infrastructure.

Building a Pipeline

A pipeline in Opsbox is a series of plugins executed in a defined sequence to produce desired outcomes. Each pipeline consists of three main components:

1. Input Plugins

These are the primary inputs/checks you want to run. They form the foundation of your analysis and are specified as a comma-separated list in the first argument.

2. Assistants

These plugins process the results from the input results. They are executed in left-to-right order, refining and expanding upon the data.

3. Outputs

Outputs handle the final results, directing them to their intended destinations. These are also specified as a comma-separated list.

Pipeline Definition Format

Pipelines are defined using the following format:

check_1,check_2-assistant_1-assistant_2-output_1,output_2

Example Pipeline

Suppose we want to analyze idle EC2 instances, have an OpenAI assistant provide cost-saving recommendations, and output the results to Jira and a text file. Our pipeline would look like this:

idle_instances-cost_savings-text_file,jira

In this example:

• idle_instances is our Rego check input.
• cost_savings is our cost-analyzing assistant.
• text_file,jira are our output destinations.

With Opsbox’s flexible plugin system, you can create powerful, automated pipelines tailored to your cloud infrastructure needs! 🚀