MetricsHub
MetricsHub Community Connectors 1.0.07
-
Home
- Connector Design Guide
Connector Design Guide
Introduction
This guide is designed to empower developers with the knowledge and best practices necessary to create effective connectors for MetricsHub, leveraging OpenTelemetry to monitor resources efficiently.
Target Audience
This guide is intended for advanced users of MetricsHub and contributors to the MetricsHub Community Connectors[1] project. Readers are expected to have a foundational understanding of monitoring concepts and YAML syntax. The content is tailored for those looking to extend MetricsHub's monitoring capabilities to resources not covered by its built-in connectors.
What is a Connector?
A connector is a YAML
file that defines the necessary parameters and logic to monitor specific resources, whether in infrastructure or any environment reachable via the network. It specifies how to collect data, what data to collect, and how to process that data. Through connectors, MetricsHub can be extended to support new resources, enriching the monitoring landscape.
Purpose of a Connector
The primary purposes of a connector are to:
- Accurately identify and detect the resources to be monitored, such as Operating Systems, Hosts, Devices, Containers, Kubernetes clusters, and more.
- Outline the techniques for data collection from these resources, utilizing local commands, network queries, or other relevant methods.
- Establish the format of the gathered data, including metrics and attributes, and detail the processes for transforming this data prior to submission to OpenTelemetry.
Key Components of a Connector
A connector YAML
file comprises several sections, each serving a distinct purpose:
- Metadata: Includes information like display name, version, platforms, and dependencies.
- Detection: Defines how to identify the target resource and includes criteria for detection.
- Monitors: Details the monitoring jobs, their types, and how they collect or compute data.
- Sources and Computes: Describe the data sources and computations performed on the collected data.
- Mapping: Describes how to create attributes and metrics
Developing a Connector
The process of developing a connector encompasses several key steps:
- Resource Identification: Determining the specific resource and its components you intend to monitor.
- Protocol Selection: Opting for the appropriate data collection method from a range of supported protocols such as
HTTP
,SNMP
,IPMI
,WinRM
,WMI
,WBEM
, andSSH
. - Connector Configuration: Creating the YAML file to specify the detection logic, metrics to be collected, and the data processing approach.
By the end of this guide, developers will be equipped with the knowledge to create connectors for MetricsHub, enabling the monitoring of a wider range of resources and contributing to the MetricsHub community.