Pipeline intelligence

Pipeline intelligence is the view into how healthy your data pipeline is: whether ingestion, data mapping, and transform jobs ran, how long they took, and whether any failed. When personalization looks wrong — stale data, missing fields, outdated recommendations — this is where you check whether the underlying data pipeline, not the personalization logic, is the cause.

What it shows

What pipeline monitoring covers
Area	What it tells you
Job status	Whether each ingestion, data mapping, or transform job is running, finished, or failed.
Run history	When each job last ran, how long it took, and how that compares to previous runs — useful for spotting a job that is slowing down or has stopped running on schedule.
Error visibility	Which job failed and why, so you know whether to fix the source connector, the mapping, or the transform logic rather than the personalization configuration.

Use it to troubleshoot

Checklist

Check job status first — if ingestion or a transform job has not completed, downstream data (and personalization results) will be stale or incomplete.
Compare the current run against run history — a job that is taking much longer than usual, or has stopped appearing, points to an upstream data or volume problem.
Read the error detail on a failed run before changing anything in Personalization — most "personalization isn't working" issues trace back to a pipeline job, not a personalization setting.
Re-run or fix the failing job at its source (connector, mapping, or transform), then confirm the next scheduled run succeeds before re-checking personalization behavior.

Discussion

Comments are not configured for local development. They appear once the site owner connects GitHub Discussions (Giscus).