pa_scores Table — ServiceNow Reference

Table Reference/pa_scores

The pa_scores table stores historical KPI values captured by Performance Analytics jobs. Part of the Performance Analytics module, this table grows extremely large on mature instances — you must always filter by indicator and date range to avoid performance issues.

What This Table Is

The pa_scores table is the data warehouse for Performance Analytics, storing every KPI value calculated by PA collection jobs. Each record represents a single score for a specific indicator at a point in time, along with dimensional breakdowns (by assignment group, category, location, etc.). This is where PA stores the actual numbers that appear in dashboards, widgets, and analytics hubs.

The table is owned by the Performance Analytics module and supports the entire PA ecosystem — from data collection through visualization. PA jobs run on schedule, calculate indicator values, and write millions of score records. These scores power PA dashboards, timeline widgets, scorecard breakdowns, and trend analysis across the platform.

The pa_scores table doesn't extend any other table — it's a standalone entity designed for time-series data. No other tables extend it either. The closest relationships are with pa_indicators (which define what to measure) and pa_cubes (which define dimensional breakdowns).

On mature instances, this table contains millions to hundreds of millions of records. Large enterprises typically see 50M+ records after two years of PA usage. The table grows continuously as PA jobs run — daily indicators add ~365 records per year per breakdown dimension. Unfiltered queries will timeout or crash your instance.

When You'll Script Against This Table

You'll script against pa_scores primarily in Scheduled Script Execution for custom reporting, Script Includes for dashboard APIs, and Business Rules for score-based notifications. Most custom PA integrations hit this table to extract historical trends, calculate custom aggregations, or export data for external analytics platforms. The pa_reader role is required for read access.

Performance Analytics has tight ACLs — only users with pa_admin or pa_power_user can modify scores. Most scripting happens in global scope since PA is typically a platform-wide capability. Background scripts inherit the running user's PA permissions, so test with appropriate role grants.

Common scripting patterns:

Extract time-series data for specific indicators to build custom charts or reports
Calculate rolling averages, percentile rankings, or year-over-year comparisons
Trigger notifications when scores cross thresholds or show concerning trends
Aggregate scores across multiple indicators for executive dashboards
Export PA data to external business intelligence tools or data lakes
Validate or correct score data after PA job failures or configuration changes
Build custom PA widgets that combine multiple indicators or apply business logic

Table Gotchas

⚠️

Never query pa_scores without filtering by indicator and date range. Unfiltered queries on large instances will timeout and potentially impact system performance for all users.

The value field stores numbers as strings — use parseFloat() for mathematical operations or comparisons
Dimensional breakdown fields (cube_1, cube_2, etc.) contain sys_ids, not display values — join to the appropriate reference table for readable names

⚠️

Score records can be updated or deleted when PA jobs re-run historical data. Don't cache score sys_ids or assume records are immutable.

The time_span field indicates the aggregation period (daily, weekly, monthly) — filter by this when you need specific granularity
Queries by date range perform best using time_span_start and time_span_end fields, which are indexed and optimized for range queries
Some indicators produce null or empty values — check value field existence before processing to avoid script errors

Free Newsletter

Enjoying this? Get one deep-dive per week.

Join 1,000+ ServiceNow pros — scripts, GlideRecord patterns, Flow Designer techniques, and career moves. Free.

The pa_indicators table defines what each score measures — indicator name, description, calculation method, and target values. Every score record references an indicator through the indicator field. The pa_cubes table defines dimensional breakdowns — how scores are segmented by assignment group, location, priority, etc.

When building queries that need readable breakdown names, you'll frequently join pa_scores with sys_user_group (for assignment group breakdowns), cmn_location (for location breakdowns), or cmn_department (for department breakdowns). The specific tables depend on which cube dimensions are configured for each indicator.

The pa_dashboards and pa_widgets tables reference pa_scores indirectly through indicator configuration. When troubleshooting widget display issues or building custom analytics interfaces, you'll often query across all three tables to understand how scores flow into visualizations.

Chrome Extension

Sourdough: ServiceNow Monitoring and Analytics

A Chrome extension for ServiceNow Admins and Developers with essential tools, analytics, graphs and monitoring features.

Instance HealthGraphs & ChartsAPI HealthDeveloper ToolsQuick SearchInstance Switcher

Add to Chrome

Free to install. Pro $5/month after a 14-day no-card trial.
Pro requires the ServiceNow admin role. Upgrade inside the extension.

Overview

Tasks

CMDB

API

Metrics

Monitor

Internals

Instance:sourdoughdev·Version:Yokohama

Instance StateONLINE

System StatusFully Operational

Session Timeout90 minutes

Logged-In Sessions2 (20 active)

Build Nameyokohama-12-18-2024_p1

IP Address10.159.128.43

Instance HealthHealth Score: 90%

🔥 5dSourdough (Chrome Plugin)Dark Mode

Key Fields

Column	Type	Label	Notes
indicator	reference	Indicator	References pa_indicators table. Always filter by this field — it's the primary key for partitioning the massive dataset.
value	string	Value	The actual KPI value as a string. Use parseFloat() for numeric operations since it's not stored as a number type.
time_span_start	glide_date_time	Time Span Start	Start of the measurement period. Indexed field — use for date range queries instead of sys_created_on for better performance.
time_span_end	glide_date_time	Time Span End	End of the measurement period. Combine with time_span_start for precise date filtering in large datasets.
time_span	choice	Time Span	Aggregation period: daily, weekly, monthly, quarterly, yearly. Filter by this when you need specific granularity for trending.
cube_1	string	Cube 1	First dimensional breakdown (often assignment group sys_id). Contains sys_id values, not display names — join to reference table.
cube_2	string	Cube 2	Second dimensional breakdown (often category or priority). Like cube_1, stores sys_ids requiring lookup for readable values.
cube_3	string	Cube 3	Third dimensional breakdown. Additional breakdown dimension — check indicator configuration to understand what it represents.
cube_4	string	Cube 4	Fourth dimensional breakdown. Most indicators use 1-3 cubes, but complex breakdowns can go to cube_10.
score_state	choice	Score State	Values: collected, manual_override. Most scores are 'collected' — 'manual_override' indicates human adjustment of calculated value.
data_freshness	choice	Data Freshness	Indicates if this score represents the most current data or historical snapshot. Critical for real-time dashboard accuracy.
target	string	Target	Target value for this KPI measurement period. Like value field, stored as string despite being numeric data.
threshold_red	string	Red Threshold	Red threshold boundary for this score. Used by PA widgets to color-code performance against targets.
threshold_yellow	string	Yellow Threshold	Yellow threshold boundary. Widgets show yellow when value falls between yellow and red thresholds.
collection_time	glide_date_time	Collection Time	When PA job calculated this score. Different from time_span_start — this is the processing timestamp, not measurement period.
anomaly_score	decimal	Anomaly Score	Statistical deviation from expected value. PA uses this for anomaly detection and automated alerting on unusual patterns.
score_uuid	string	Score UUID	Unique identifier for this specific score calculation. PA uses this to prevent duplicate processing when jobs retry.
snapshot_id	reference	Snapshot ID	References pa_snapshots if this score is part of a snapshot collection. Null for regular time-series scoring.

Common Queries

Get all scores for a specific indicator in the last 30 days

// Basic query pattern — always filter by indicator and date
var gr = new GlideRecord('pa_scores');
gr.addQuery('indicator', 'your_indicator_sys_id');
var thirtyDaysAgo = new GlideDateTime();
thirtyDaysAgo.addDaysUTC(-30);
gr.addQuery('time_span_start', '>=', thirtyDaysAgo);
gr.orderBy('time_span_start');
gr.query();
while (gr.next()) {
    gs.info('Score: ' + gr.getValue('value') + ' for period ' + gr.getDisplayValue('time_span_start'));
}

Get scores for specific indicator filtered by assignment group breakdown

// Filter by indicator, date range, and dimensional breakdown
var gr = new GlideRecord('pa_scores');
gr.addQuery('indicator', 'incident_resolution_time_indicator_id');
gr.addQuery('cube_1', 'assignment_group_sys_id'); // cube_1 often contains assignment group
var lastQuarter = new GlideDateTime();
lastQuarter.addMonthsUTC(-3);
gr.addQuery('time_span_start', '>=', lastQuarter);
gr.addQuery('score_state', 'collected'); // Exclude manual overrides
gr.orderByDesc('time_span_start');
gr.query();
while (gr.next()) {
    gs.info('Group score: ' + parseFloat(gr.getValue('value')).toFixed(2) + ' minutes');
}

Find indicators with scores exceeding red threshold this month

// Real-world scenario: identify KPIs performing below target
var gr = new GlideRecord('pa_scores');
var thisMonth = new GlideDateTime();
thisMonth.setDayOfMonthUTC(1); // First day of current month
gr.addQuery('time_span_start', '>=', thisMonth);
gr.addQuery('time_span', 'daily');
gr.addQuery('value', '>', 'threshold_red'); // Value worse than red threshold
gr.addNotNullQuery('threshold_red');
gr.query();
while (gr.next()) {
    var indicator = gr.getDisplayValue('indicator');
    gs.info('Alert: ' + indicator + ' scored ' + gr.getValue('value') + ' (red threshold: ' + gr.getValue('threshold_red') + ')');
}

Count scores by indicator for the current year using GlideAggregate

// Aggregate pattern for counting score records by indicator
var ga = new GlideAggregate('pa_scores');
var yearStart = new GlideDateTime();
yearStart.setDayOfYearUTC(1); // January 1st of current year
ga.addQuery('time_span_start', '>=', yearStart);
ga.addAggregate('COUNT');
ga.groupBy('indicator');
ga.orderByAggregate('COUNT');
ga.query();
while (ga.next()) {
    var count = ga.getAggregate('COUNT');
    var indicator_name = ga.getDisplayValue('indicator');
    gs.info(indicator_name + ': ' + count + ' score records this year');
}

Wrong way vs right way to query by date range

// WRONG: Using sys_created_on is slow and unreliable
// var gr = new GlideRecord('pa_scores');
// gr.addQuery('sys_created_on', '>=', lastWeek);

// RIGHT: Use time_span_start for date filtering
var gr = new GlideRecord('pa_scores');
gr.addQuery('indicator', 'required_indicator_sys_id'); // Always filter by indicator first
var lastWeek = new GlideDateTime();
lastWeek.addDaysUTC(-7);
gr.addQuery('time_span_start', '>=', lastWeek); // Indexed field for measurement period
gr.query();
// time_span_start represents when the measurement occurred, not when the record was created

Related Tables

pa_indicators sys_user_group

Test Your Knowledge

Quick 3-question quiz — see how your ServiceNow skills stack up.

Question 1 of 3Performance

A list view on a table with millions of records is slow. Best fix?

Select an answer to continue