Data Management
Work queues
Work queues are the backbone of scalable automation. Items have a key, data, priority, status, and tags. Statuses include Pending, In Progress, Completed, Exceptioned, and Deferred. Understand deferral (retry later), priority, and tags for filtering and reporting.
Collections and data items
Collections hold tabular data in memory. Understand collection fields, filtering, and the difference between a collection and a queue: collections are transient in-process data, queues are persistent and shared across workers.
Traps
A frequent trap is using a queue as a general data store rather than for work items. Another is misunderstanding item locking: when a worker gets the next item, it is locked to that resource until completed or the lock times out. Questions may test what happens when a worker crashes mid-item: the item returns to the queue after the lock expires, so idempotency matters. Also know that queue data can be encrypted.
Data quality
Validate inputs, handle duplicates via keys, and archive completed items according to retention policy. Avoid storing sensitive data in queue data unless encrypted.
Rule of thumb
Design every queue item to be safely retried: if it runs twice, the outcome should be the same.
Worked example
An invoice queue uses the invoice number as the key to prevent duplicates. A worker locks an item, posts to the finance system, and marks it complete. If it crashes after posting, the retry must check whether the invoice already exists before posting again.