scout

In the Scout Announcement blog post I mentioned we are looking to integrate Goss into Scout and I wanted to post an update on that.

Background

Goss is something similar to serverspec - it lets you write unit tests about your nodes actual state rather than code used to build it. Goss definitions are written in YAML or JSON and supports Go templating for customization.

This model is well suited for the purposes of monitoring since you can write really in depth sets of validations and treat them as a single unit.

Goss is written in Go, very fast and thanks to a lot of work I did recently embeddable in other software.

Yesterday I introduced a new Choria component called Scout which helps you build scalable monitoring pipelines. Today, we’ll look a bit at what makes a Scout install and how it is built.

In a follow up post I’ll dive a bit into Autonomous Agents - an infrequently used but very powerful building block found in Choria.

Overview

We’re happy to announce a new project called Choria Scout - a highly scalable system health monitoring framework and monitoring data pipeline released under the Apache 2.0 license.

Initially we support the ability to execute Nagios compatible plugins on Choria managed nodes with results sent centrally in a standard CloudEvents format, and optionally, integrated into Prometheus.

These are framework level building blocks that will in time be used to create a full monitoring stack built on Choria technologies. Checks and value overrides can already be configured using our Puppet modules. You can also use these building blocks to build entirely custom solutions for your own needs.

Scout will be a cloud native project with central components capable of being hosted on Kubernetes and using data formats supported by commercial clouds and projects like KNative. It will have a focus on integration, open data exchange and extensibility.

Despite being cloud native we will of course support monitoring anything where Choria, or the upcoming Scout agent, can run which includes traditional baremetal, VMs, containers and pods and small devices.