Introduction to Entities and Metrics
New Geneos data model Copied
With the release of Gateway Hub, ITRS is taking the opportunity to streamline the data model, allowing for far simpler configuration and ease of use. From version 1.3 of Gateway Hub, the new data model is far simpler, consisting of Entities and Metrics.
Entities depend on other Entities, and each Entity has a set of Metrics:
Entities Copied
Entities represent something being monitored, and can correspond to a physical or a logical item such as:
- Servers
- Containers
- Virtual Machines
- Applications
- Processors
- Storage
The new model for Entities expands upon the Managed Entity concept in Geneos, bringing the concept of relationships between Entities, such as a Processor depending on a physical machine, or an application running in a container. This allows more advanced capacity modelling when using ITRS Capacity Planner.
As in the old data model, Entities can be assigned Attributes, which are key/value pairs used for identification and grouping. Managed Entities defined previously in the Gateway map one-to-one to the new system without issue.
Metrics Copied
A Metric is a measurement from an Entity. It has a time and a value component, and aligns with a dataview column or headline cell.
Why are we making this change? Copied
Historically, the Geneos data model was based around the technical hierarchy of the following:
This hierarchy has a number of drawbacks including:
- It forces you as a user to be aware of technical details you might not be interested in.
- It is not as suited to dynamic environments e.g. Self-Announcing Netprobes rebalancing across Gateways, containers, and orchestration.
Going forward, the simplified entity model means we can decouple the implementation details of how Geneos performs monitoring (configuration of Netprobes, Gateway, and so on), allowing you to focus on what Geneos is monitoring (departments, applications, entities and their metrics).
Although things may seem a little more complicated initially as the simplified model is bedded in, it will soon result in a substantially simpler, more intuitive model for monitoring.