what are the roles & limitations of activity chaining ? in which kind of scenarios we use activity chaining? in which kind of scenarios chaining may break generally?

You use activity chaining to provide a synchronous experience of the process to an end user. For example if you have a simple process that when you start the first node is a User Input Task (UTI. If you do not chain from the Start Node to the UIT then the Task will be assigned to the user asynchronously and they'll have to go and find that Task in a Task list. If you do chain from the Start Node to the UIT then the User Interface that is implemented in the UIT will immediately be presented to the User.

Chaining typically breaks after the chain limit that is set for the environment, which is normally 50 nodes. Note that includes individual nodes that might appear within sub-processes (i.e. a sub=-process node is not considered to be one node for chaining purposes, it includes all of the nodes contained within that that sub-process).