Spinnaker Feature – Management Flexibility: Cluster Management and Deployment Management
Apr 20, 2017 by Armory
Spinnaker ships with two core features: cluster management and deployment management. These features are easily reviewed through Armory’s Lighthouse feature.
Cluster Management – Managing your Clusters and regions can get out of hand as you scale – therefore Spinnaker strives to simplify it with an interactive and intuitive UI. Through this Spinnaker UI you can also manage the Security Group and Load Balancers associated with the respective Server Groups through Spinnaker.
<img src=”https://d1ax1i5f2y3x71.cloudfront.net/items/1u2q3W3U3v3t1J0x3o1S/Image 2017-03-24 at 3.05.04 PM.png“)
- Cluster – A logical or customized grouping of Server Groups
- Server Group – A collection of Virtual Machines (VM) running instances of images
- Security Group – Generally a firewall that determines network traffic access. Can be set to desired specifications through Spinnaker
- Load Balancer – Balances traffic among Server Group instances
Deployment Management – Deployments can take many manual steps that are hard to script and track. Spinnaker allows for precise orchestration of the Stages and Pipelines containing them to get your desired execution.
- Stage – A Stage is an action within a Pipeline; successive completion of Stages leads to the eventual execution of a Pipeline.
- Pipeline – A logical series of actions called Stages for deployments, where parameters are passed within each Pipeline from Stage to Stage. Pipelines can be configured to start manually or automatically triggered as well as send notifications and self-diagnosis to external channels such as SMS, Slack, email, etc.
Lighthouse – Spinnaker is for managing Clusters and Deployments, but Spinnaker itself should also be easy to manage. Lighthouse is a lightweight service for managing Spinnaker specific to Armory’s distribution of Enterprise Spinnaker.
- Monitors Spinnaker services for health and failures of your deployed clusters
- Monitors Orca on its own instance for termination when no jobs are running
- Enables easy upgrades for Spinnaker (Spinnaker is an amalgamation of multiple services for which the user would have to manually update)
- Bootstraps Spinnaker with pipelines that can be used in integration testing
- Aggregates Spinnaker’s metrics for displaying onto a dashboard
all of the related posts here.
