Or several. It depends on how good of a model you want of a 'real' web server with separate web server and database, plus separate test system to drive the load. I wouldn't spend money for it from scratch, but if you already have recent hardware in a small form factor it would be worth considering.
Log4j supports selective logging. That means you can have info/debug/trace priority messages in place, but never see them in the log unless you explicitly enable extra logging for that class or package. You can do this at runtime, e.g., via something like 'chainsaw' (which attaches to a running process) or hooks in your UI.
Our policy is that logs are usually very quiet. Application startup/shutdown and not much more. But if there's a problem the debugging messages are already in place to let us peak into the system, even if it's been deployed to a production site.
BTW AOP is also great for this. You can configure logging interceptors that log activity when you're in a development environmnent, but easily removed in production. This is a natural approach when going from one layer to the next, e.g., when wrapping the DAO layer.