PeopleSoft Performance Monitor: Complete Guide & Practical Uses

Master PeopleSoft Performance Monitor (PPM): set up monitoring systems, understand PMUs and Events, configure Agent Filter Levels, troubleshoot slow transactions, manage data volume, and use Performance Trace. Step-by-step walkthrough of every PPM feature with real examples. For PeopleSoft administrators, DBAs, and IT professionals.

🎯 Essential PeopleSoft Skill: Performance Monitor is the primary diagnostic tool for the majority of PeopleSoft performance issues. Monitor transactions, identify bottlenecks, troubleshoot slow sign-ons—all from the PeopleSoft console. This guide covers every feature with step-by-step instructions and real-world examples. Beginners to advanced administrators.

📋 PeopleSoft Performance Monitor Complete Guide

1. PPM Overview & Architecture
2. Setting Up the Monitoring System
3. Setting Up the Monitored System
4. Understanding PMUs and Events
5. Agent Filter Levels Explained
6. Performance Trace: Targeted Troubleshooting
7. Performance Monitor Overhead
8. Data Volume Management
9. Troubleshooting Common Performance Issues
10. Pro Tips & Console Tricks

What is PeopleSoft Performance Monitor? Built-in diagnostic framework in PeopleTools (available since 8.44) that tracks performance across every tier of the PeopleSoft Pure Internet Architecture—web servers, application servers, process schedulers, and the database. Collects transaction data, server events, and timing information, all without writing custom code.

How PPM Works: (1) PPM Agents run inside each PeopleSoft process (web server, app server, process scheduler). (2) Agents collect performance data and buffer it locally. (3) Agents send data via HTTP to the PPMI Monitor Servlet on the monitoring system’s web server. (4) The PSPPMSRV collator process receives the data and stores it in the PPM database. (5) Administrators query this data through the Performance Monitor analytics pages.

Two-System Architecture: PPM requires a separate monitoring system—the system collecting performance data must be completely separate from the systems being monitored. Oracle does not support self-monitoring configurations. The monitoring system needs its own database, web server, application server, and process scheduler.

Main PPM Sections: System Monitor (live performance view), Analytics (User Requests, Top Components, Component Trace), History (Completed PMUs, Event History, User Sessions), Administration (setup, agent filters, scheduling), System Performance (real-time dashboard).

Best for Administrators: PPM console is the most practical way to diagnose production issues without impacting users. Use it to pinpoint slow transactions, identify resource bottlenecks, and track SQL performance directly from the browser.

Step 1: Create the Monitoring Database. Create or upgrade a PeopleTools System database. Application DDL scripts aren’t needed—only PeopleTools tables are used. Use the Database Configuration Wizard and select Database Create Type = “PeopleTools System.” If cloning an existing database, set the GUID column in PSOPTIONS to a space before use.

Step 2: Create the Application Server Domain. Configure with at least 2 PSAPPSRV processes. In the Quick Setup menu, enable option “10. Perf Collator.” This tells Tuxedo to start a PSPPMSRV process—the collator that receives and stores all performance data.

Step 3: Deploy PIA and Create a Process Scheduler. During PIA setup, record the web site name, hostname, port number, and web profile name. Create a process scheduler domain to run the Archive, Reaper, and Lookup Application Engine programs.

Step 4: Set Up PPMI User Security. Navigate to PeopleTools, Security, User Profiles. Create a dedicated user (e.g., PPMAdmin), assign the “PeopleTools Perfmon Client” role, and confirm the PTPMCLNT permission list has the Performance Monitor PPMI Access checkbox selected.

Step 5: Configure the Integration Broker Gateway and Filter Level. Enter the gateway URL at PeopleTools, Integration Broker, Configuration, Gateways. Then navigate to Global Administration, System Defaults and confirm the Agent Filter Level is set to 04-Standard. Always click Save before Apply to Current Systems.

Step 6: Schedule the Maintenance Programs. PSPM_REAPER and PSPM_LOOKUP every 15 minutes (PerfMon Reaper Recurrence), and PSPMARCH daily at 1:00 AM (Daily Purge recurrence) with Run %UpdateStats checked.

Grant PPM Admin Access: Link the admin User ID to the PTPT1200 permission list (or a custom equivalent) to access Performance Monitoring menu items.

Enable the PPM Agent: In psappsrv.cfg and psprcs.cfg, under the [PSTOOLS] section, set EnablePPM Agent=1. Restart the domain after the change.

Verify the Web Profile: Confirm “Enable PPM Agent” is checked and PPM Monitor Buffer Size is at least 51,200 KB. Restart the web server if changed.

Specify the Monitor URL: Navigate to PeopleTools, Performance Monitor, Administration, Specify Monitor. Enter the monitoring system URL ending with a forward slash, e.g., http://host:port/monitor/ps/. Click Save and Ping Test to verify connectivity.

Reboot and Register: Log into the monitored system to trigger agent registration, then wait one agent heartbeat interval (default 5 minutes) for full registration.

Performance Measurement Units (PMUs): A PMU captures the execution of a specific section of PeopleSoft code, with start/stop times, duration, and metrics. PMUs form parent-child trees showing the full hierarchy of a user request—from the browser through the web server, application server, down to individual SQL executions.

Key PMU Types: PMU 100 (Portal Request) and 101 (PIA Request) require Standard level. PMU 108 (sign off) and 109 (log on) are always collected regardless of filter level. PMU 115 (JOLT Request) and 400 (Tuxedo Service PCode and SQL) are Standard. PMUs 406-408 (SQL execution variants) require Verbose. PMU 404 (PeopleCode Program Execution) requires Debug.

Events: Point-in-time notifications about resource usage—CPU, JVM memory, JOLT exceptions, server recycles. Unlike PMUs, events have no duration. Key warning events: 355 (query killed), 356 (server recycle/shutdown), 500 (JOLT service exception), 801 (PPMI Monitor Servlet buffer overrun), 900 (agent buffer overrun).

Performance Data Attributes: Each PMU has Context fields (common to the entire request—component name, session ID, IP address) and Metric fields (measurements taken during the PMU’s lifetime—SQL fetch counts, buffer sizes, timing values).

What is the Agent Filter Level? Controls how much performance data PPM collects. Higher levels capture more detail but generate more data and slightly higher overhead. Set at the system level—individual agent types can override it.

The Six Levels: 01-Standby (agents active, no data sent), 02-Error (error data only), 03-Warning (warning events like JOLT exceptions and server recycles), 04-Standard (PIA/Portal requests, sign-on/off—recommended for production), 05-Verbose (adds SQL performance data), 06-Debug (adds PeopleCode execution statistics).

Changing Filter Levels: Navigate to PeopleTools, Performance Monitor, Administration, Agent Filters. Select the monitored system, adjust the filter level per agent type (PSAPPSRV, PSMONITORSRV, WEBRESOURCE, WEBSERVER), then click Save and Notify Agents.

Production Warning: Never set 05-Verbose or 06-Debug system-wide in production. Network bandwidth at Verbose level is roughly 7x higher than Standard. Use Performance Trace instead for targeted, user-specific debugging.

What is Performance Trace? Lets a single user activate Verbose or Debug level monitoring for their own session only, without impacting any other user on the system. The recommended way to collect detailed data in production.

Admin Setup: (1) Create a permission list with full access to the WEBLIB_PPM web library. (2) Add the permission list to a role (e.g., “PPM Console Access”). (3) Assign the role to users who need trace access.

How Users Activate a Trace: A “Performance Trace” link appears in the portal header. Click it, name the trace, select a filter level (Standard, Verbose, Debug), click Start Trace, reproduce the issue, then click Stop Trace.

Reviewing Results: Navigate to PeopleTools, Performance Monitor, Analytics, User Requests and select the trace name. For deeper analysis, use Analytics, Component Trace—run at minimum 05-Verbose to see the SQL Summary, or 06-Debug to see the PeopleCode Summary.

CPU Overhead (1,000 concurrent user benchmark): PPM Disabled = 38.20% app server CPU / 7.48% web server CPU. PPM on Standby = 38.25% / 7.62% (negligible impact). PPM on Standard = 39.75% / 8.01% (less than 2% increase on the app server, less than 0.5% on the web server).

Network Bandwidth: At 04-Standard, roughly 720 Kbits/second. At 05-Verbose, this jumps to roughly 4,800 Kbits/second—nearly 7x higher, which is why Verbose should never run system-wide in production.

Memory Overhead Formulas: App Server = Max Buffer Size × (Number of PSAPPSRV processes + 1). Web Server = Max Buffer Size × (Number of Sites + 1). Process Scheduler = Max Buffer Size × (Master Scheduler + 1). Default Max Buffer Size is 4MB (4,194,304 bytes).

Database Overhead: In the 1,000-user benchmark, roughly 740,000 PMU rows and 2,000 event rows were generated, totaling approximately 2,015 MB of database space.

Why It Matters: At 04-Standard, PPM generates 5-7 PMUs per PIA request and 7-9 PMUs per Portal request, each occupying 2.4-3.8 KB in a Unicode database. Tables grow quickly without proper maintenance.

The Reaper Program: Runs every 15 minutes via PerfMon Reaper Recurrence. Deletes flagged rows in PSPMTRANSCURR, marks expired PMUs as timed out, and moves them to the PSPMTRANSHIST history table.

Archiving: PSPMARCH moves data older than your retention threshold from history tables to archive tables. PeopleSoft recommends archiving data older than 7 days in production. Archive modes: Delete Data, Archive Data, Archive Nothing, or Delete System.

PMU Sampling: Set the Agent PMU Sample Rate (1/X) to a positive integer N—only 1 in N requests generates PMUs. Sign-on PMUs (108, 109, 116) are always captured regardless of sampling.

Collator Row Limit: Set in Global Administration; default is 0 (unlimited). When the limit is reached, PSPPMSRV stops inserting data and all monitored systems appear Stale. Recommendation: leave at 0 unless you have a specific operational reason to cap it.

Issue 1: Slow Sign-On. Check: (1) Multiple pagelets—search PMU 100 where PMU Details = Homepage and drill into GetContent PMUs for slow ones. (2) Slow authentication—search PMU 101 where Detail = Login and inspect Authenticate/GetCertificate timing. (3) Network latency—check Current User Sessions Details tab or have the user run PeopleSoft Ping.

Issue 2: Long-Running Requests. Navigate to Analytics, User Requests, enter the User ID and date range, then sort by Duration. In the User Request Details, high SQL Total Time suggests database tuning opportunities; high Serialization Time suggests an oversized Component buffer; high Cache Misses are normal after a domain restart but concerning if persistent.

Issue 3: “Stale Agent Data Detected.” Check in order: setup steps completed correctly, Agent Filter Level isn’t 01-Standby or Off, the domain isn’t inactive, a web domain may simply have no logged-in users yet, and the Collator Row Limit hasn’t been reached.

Issue 4: PerfMon Agent Error 500 or 404. Error 500 means the agent can’t connect to the PPMI Monitor Servlet—check the site name in the Specify Monitor URL. Error 404 (“Monitor not activated”) means the PSPPMSRV collator isn’t running—verify the Perf Collator option is enabled on the monitoring application server domain.

Issue 5: No Performance Data Collected. Common causes: incorrect permission lists, wrong URL values, Perf Collator not enabled, Agent Filter Level set to 01-Standby, Collator Row Limit reached, or the Agent PMU Sample Rate set too high.

PeopleSoft Performance Monitor Training & Administration Services

PepperTech offers PeopleSoft administration training and consulting covering PPM setup, performance tuning, PeopleTools upgrades, and production support. Expert trainers with 15+ years of PeopleSoft infrastructure experience. Get your team productive with PPM in days, not weeks.