Four building blocks
web or eu-west. Steps and variables can be scoped to tags.
Environments are scopes
When you deploy a release to Production, Jaws Deploy resolves variables for Production, runs the steps that apply to Production and records the deployment in Production's history. A lifecycle decides which environments a release has to pass through first.
Use environments for stages and tags for groups of servers. With one environment per server, variable scoping and lifecycles quickly become hard to manage.
Machines and the agent
The agent runs as a Windows service or a Linux daemon and opens an outbound WebSocket connection to Jaws Deploy. Servers behind NAT or a strict firewall only need to reach Jaws Deploy, and an administrator approves each new agent before it gets any work. Install one agent per machine. Servers without internet access can install the agent offline.
Tag conventions
Tags are free text, so agree on a naming convention early. For example:
web # application servers
db # database hosts
cache # Redis
eu-west # region
large # capacity tier
When to use a machine
Windows services, IIS sites and anything else that has to run on the server itself. The agent runs the steps there.
When to use a cloud target
Azure Web Apps, where Azure runs the host and there is no agent to maintain. One project can deploy to machines and cloud targets.
A typical setup
Three environments (Development, Staging and Production), each with a few tagged machines and an Azure Web App. Variables are scoped to an environment, a tag or a single target, and no step names a specific machine.
To add a server, install the agent, approve it and give it the right environment and tags. Existing projects deploy to it on their next deployment.