Data Collection
About 22% of the CIS-PA exam — see all 6 domains.
Data collection is the engine that turns indicator and breakdown definitions into stored scores and snapshots. Understanding the job, what it produces, and how to operate it is essential.
Data collection jobs
A data collection job runs on a schedule and collects scores for the indicators (and their associated breakdowns) that it is configured to process. The default behavior collects scores once per period (commonly daily). Jobs can be scheduled and can also be run on demand for testing or backfilling. Each run produces:
- Scores — one value per indicator per period, and one per breakdown element where breakdowns apply. Scores are stored in the PA score tables and form the time series.
- Snapshots — a snapshot is the list of record
sys_ids that were counted/aggregated to produce a score at collection time. Snapshots are what let you click a score on a scorecard and drill through to the exact underlying records as they existed then. Snapshots are made for indicators (subject to the "collect records for this indicator" setting).
Historic data collection
You can collect historic data to backfill scores for past periods using the data currently in the tables (an approximation, since it uses current values). This is how a new indicator gets a usable trend line immediately instead of waiting weeks for history to accumulate. Be aware historic collection is an estimate and can be resource-intensive.
Operating and troubleshooting collection
- Job logs record what each run collected and surface errors; reviewing the data collection job log is the first step in troubleshooting missing scores.
- Common causes of missing or wrong scores: the indicator source condition excludes the records, the breakdown matrix entry is missing, the job did not run, or the frequency is misconfigured.
- Data cleanup / retention jobs remove old scores and snapshots per retention settings to control storage.
- Performance considerations: limit unnecessary breakdowns, reuse indicator sources, and stagger jobs.
A frequently tested chain of cause and effect: indicator source defines records → data collection job runs on a frequency → produces scores (time series) and snapshots → scores feed widgets/scorecards → snapshots enable drill-through.
Sample questions from this domain
Three of the 64 in this domain, with the reasoning. The full set is in the question bank.
What is the primary function of a Performance Analytics data collection job?
- A. To send email notifications to users
- B. To create incidents automatically
- C. To run on a schedule and collect indicator scores into the scores table ✓
- D. To build Service Portal pages
Why: A data collection job executes on a defined schedule to aggregate records and write indicator (and breakdown) scores to the scores table. It does not send notifications, build portal pages, or create incidents.
Where are collected indicator scores stored?
- A. In the incident table
- B. In the email log
- C. In the scores table (pa_scores) ✓
- D. In the sys_report table
Why: Scores produced by collection are stored in the pa_scores (scores) table, which holds the time series. They are not stored in sys_report, the incident table, or the email log.
An administrator wants to backfill scores for the past 12 months for a newly created indicator. Which type of collection should be run?
- A. Historical data collection ✓
- B. Real-time report refresh
- C. Catalog order guide
- D. Discovery schedule
Why: Historical data collection recalculates scores for past periods, backfilling the trend so the indicator has meaningful history. A real-time report refresh, catalog order guide, and Discovery schedule are unrelated to backfilling PA scores.