Real-Time Indexing Backlog

This alert triggers when a record’s Real-Time Indexing queue in PTSF_RTISTG holds too many index updates or its oldest update has waited too long.
Categories:
New to psLens? This page documents one specific alert. To see how it appears on the dashboard, what operators investigate, and how teams tune it, start with a live walkthrough.
Book a Demo Browse Alerts
Tailored Operational Context
  • Target Database: —
  • Context Type: —
  • Alert Severity: —
  • Triggered Time: —
  • Firing Context: —
  • Contextual SQL: 
    —

Real-Time Indexing Backlog Alert

Alert ID: rti_staging_backlog Category: Search Framework Default thresholds: 1,000 queued updates, or 30 minutes Runs: every 15 minutes

What This Alert Detects

With Real-Time Indexing on, a database trigger on each indexed record writes a row to the queue table PTSF_RTISTG for every insert, update, and delete. PeopleSoft then sends those rows to the search engine and removes them. This alert triggers when the queue for a record is not draining.

Two situations cause it:

  • The process that drains the queue has stopped. Rows keep arriving and none leave, so the oldest row gets older.
  • A bulk update or batch load changed thousands of rows on a record that carries a Real-Time Indexing trigger. The queue fills faster than it drains.

Severity Logic

ConditionSeverity
A record has more than thresholdCount queued updatesWarning
A record’s oldest queued update is thresholdMinutes oldWarning

Either condition triggers the alert. There is one alert item per record.

What Gets Checked

The alert runs one aggregate query on PTSF_RTISTG, grouped by RECNAME, returning the row count and the oldest PTRTIINSERTDTTM per record. The age is measured against the database server’s clock.

Alert Details

Each alert item includes:

  • The record whose trigger queued the updates (RECNAME)
  • The number of queued updates
  • When the oldest update was queued, and how many minutes ago
  • A link to the Search Framework Monitor

Configuration

alerts:
  checks:
    rti_staging_backlog:
      enabled: true
      thresholdCount: 1000    # Queued updates per record before alerting
      thresholdMinutes: 30    # Age of the oldest queued update before alerting
SettingDefaultDescription
thresholdCount1000Queued updates on one record that trigger the alert
thresholdMinutes30Age in minutes of the oldest queued update that triggers it

How to Respond

  1. Open the Search Framework Monitor and switch on the Real-Time Indexing Queue panel to see the oldest queued rows and their keys.
  2. If the count is high and the oldest row is recent, a bulk change is in progress. Check Process Monitor for the batch job, and confirm the queue shrinks once it finishes.
  3. If the oldest row keeps getting older, the queue is not draining. Check that the Real-Time Indexing server process is running in the application server domain and that the search engine is reachable.
  4. Open the record’s detail page and switch on Search Framework Usage to see which search definitions index it.

Required Table

PTSF_RTISTG. Without it the alert is inactive.