Worker and Scheduler both support a {Worker|Scheduler}.story} method which filters the transition log and returns all events for a given key or stimulus ID. Often, to understand the story the entire cluster-wide history is relevant.

The Client should support an API returning the story of the a key or stimulus ID.

• It must be clear where the event originated from (That's important to distinguish clocks)
• The failure behaviour of timed out workers should be configurable (raise, ignore). TBD: Should there be a sentinel if ignored?

Note: The log format for scheduler and worker events are still different.

Stimulus IDs on scheduler side only supported once #5849 is implemented