Background Script: Tag incident resolution outliers using z score (#2466)

Author: franks883

Server-Side Components/Background Scripts/Tag Incident Outliers/README.md

@@ -0,0 +1,25 @@
+# Tag incident resolution outliers by z score
+
+## What this solves
+Average resolution time hides long-tail outliers. This script calculates mean and standard deviation of resolution minutes and tags incidents whose z score exceeds a threshold, helping teams investigate anomalies.
+
+## Where to use
+Run as a Background Script or convert into a Scheduled Job for periodic tagging.
+
+## How it works
+- Uses `GlideAggregate` to compute count, mean, and approximate variance
+- Calculates z score per resolved incident
+- Sets a flag field or work note on outliers above a configurable z threshold
+
+## Configure
+- `DAYS`: look-back window
+- `Z_THRESHOLD`: default 2.5
+- `FLAG_FIELD`: field to set, for example a custom boolean `u_outlier`
+
+## References
+- GlideAggregate API  
+  https://www.servicenow.com/docs/bundle/zurich-api-reference/page/app-store/dev_portal/API_reference/GlideAggregate/concept/c_GlideAggregateAPI.html
+- GlideRecord API  
+  https://www.servicenow.com/docs/bundle/zurich-api-reference/page/app-store/dev_portal/API_reference/GlideRecord/concept/c_GlideRecordAPI.html
+- GlideDateTime API  
+  https://www.servicenow.com/docs/bundle/zurich-api-reference/page/app-store/dev_portal/API_reference/GlideDateTime/concept/c_GlideDateTimeAPI.html

Server-Side Components/Background Scripts/Tag Incident Outliers/tag_resolution_outliers.js

@@ -0,0 +1,60 @@
+// Background Script: Tag incident resolution outliers by z score
+(function() {
+  var TABLE = 'incident';
+  var DAYS = 30;
+  var Z_THRESHOLD = 2.5;
+  var FLAG_FIELD = 'u_outlier'; // create this boolean field or change action to add work_notes - potential to change to tag as well if one exists
+
+  // Build look-back cutoff
+  var cutoff = new GlideDateTime();
+  cutoff.addDaysUTC(-DAYS);
+
+  // First pass: mean and std dev of resolution minutes
+  // Compute duration per record as closed_at - opened_at in minutes
+  var minutes = [];
+  var gr = new GlideRecord(TABLE);
+  gr.addQuery('closed_at', '>=', cutoff);
+  gr.addQuery('state', '>=', 6); // resolved or closed
+  gr.addNotNullQuery('opened_at');
+  gr.addNotNullQuery('closed_at');
+  gr.query();
+  while (gr.next()) {
+    var opened = String(gr.getValue('opened_at'));
+    var closed = String(gr.getValue('closed_at'));
+    var mins = gs.dateDiff(opened, closed, true) / 60;
+    minutes.push({ id: gr.getUniqueValue(), mins: mins });
+  }
+  if (!minutes.length) {
+    gs.info('No records in window. Exiting.');
+    return;
+  }
+
+  var sum = minutes.reduce(function(a, x) { return a + x.mins; }, 0);
+  var mean = sum / minutes.length;
+
+  var variance = minutes.reduce(function(a, x) {
+    var d = x.mins - mean; return a + d * d;
+  }, 0) / minutes.length;
+  var std = Math.sqrt(variance);
+
+  // Second pass: tag outliers
+  var tagged = 0;
+  minutes.forEach(function(row) {
+    var z = std > 0 ? (row.mins - mean) / std : 0;
+    if (z >= Z_THRESHOLD) {
+      var r = new GlideRecord(TABLE);
+      if (r.get(row.id)) {
+        if (r.isValidField(FLAG_FIELD)) {
+          r[FLAG_FIELD] = true;
+          r.update();
+        } else {
+          r.work_notes = 'Marked outlier by automation. z=' + z.toFixed(2) + ', mean=' + Math.round(mean) + 'm, std=' + Math.round(std) + 'm';
+          r.update();
+        }
+        tagged++;
+      }
+    }
+  });
+
+  gs.info('Outlier tagging complete. Window=' + DAYS + 'd, N=' + minutes.length + ', mean=' + Math.round(mean) + 'm, std=' + Math.round(std) + 'm, tagged=' + tagged);
+})();