Data Modelling

What does the Message Feedback Event Dataset actually cover — just email bounces, or more?

All channels (Email, SMS/RCS/MMS, Direct Mail, etc.) — delivery logs plus email ISP bounce feedback specifically. It's broader than "email bounces," used for both reporting and audience creation.

What's the difference between the Journey Step Event dataset and the Journeys dataset?

Journey Step Event captures every Journey Step Experience Event generated at runtime (consumed by Reporting and CJA for YoY analysis, tied to Journey Metadata);

Journeys is the metadata dataset describing each step's configuration

— one is runtime event data, the other is journey structure/metadata.

Besides the SQL and name fields, what capability can a query template have via the /query-templates API that turns it from a reusable query into a recurring job?

A scheduler (cron expression) — this lets a query template run on a recurring schedule rather than just being triggered ad hoc.

What's the execution time limit for an ad hoc (interactive) Query Service query, vs. a batch (CTAS/ITAS) query?

Ad hoc: 10 minutes. Batch: up to 24 hours — the 10-minute ad hoc timeout does not apply to batch queries.

What's the default UI result row limit for Query Service, and the max value allowed in a LIMIT clause?

50,000 rows returned via the UI by default; LIMIT clause can go up to 100,000.

How many concurrent ad hoc Query Service user sessions are allowed per org by default, and can this be increased?

10 per org by default, expandable via an "Ad Hoc Query Users Add-On Pack."

What's the difference between a "Performance Guardrail" and a "System-enforced Guardrail" in Query Service?

Performance (soft limit): exceeding it degrades performance/latency, but isn't blocked — Adobe isn't responsible, you can license more capacity.

System-enforced (hard limit): the UI/API actively blocks the action or returns an error — cannot be exceeded.

What are the 6 cloud storage destinations available for exporting AJO datasets, and where do you configure them?

Amazon S3,
Azure Blob,
Azure Data Lake Gen2,
Data Landing Zone,
Google Cloud Storage,
SFTP
— configured via the Destinations menu, Catalog tab.

What's the actual limitation on how much historical data can be included in a dataset export?

Data older than 365 days is excluded — a rolling data-age cutoff, not a matter of picking an end date.

Can a dataset created via the API be exported to a cloud storage destination?

No — datasets created via API are not available for export, regardless of destination.

If you need to delete a dataset that's currently being exported to a destination, what must you do first, and why?

Remove the dataset from the destination's dataflow BEFORE deleting it — the UI does not block you from deleting a dataset mid-export, so skipping this step breaks the export silently.

Why can't you trust the monitoring metrics shown for dataset exports as an exact count of dataset export activity?

Because export monitoring metrics are currently mixed together with profile export numbers — the numbers shown don't isolate dataset-only export activity.

Difference between ad hoc query, query template and batch query

ad hoc = write a query as usual in query editor
query template = save the ad hoc query - can include a schedule, which creates a cron job
batch query = click “run as CTAS” (Create Table As Select) on a query template. This will create a dataset from the query