The reason to have multiple agents is if you need them due to mutually incompatible requirements (i.e. windows agents and linux agents) or because you need them to reduce wait times. Otherwise you should be able to use a single agent.
You can do it with one agent, as long as the agent can connect to the 3 servers. It depends on where you are deploying and how you are deploying.
In short, agents doesn't represent where the code is deployed it. They are the servers where the actual build is done. If your process needs the deployment to be done in the target env, then you will need as many agents.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.