Understanding the "SQL PI Experiencing Errors" Rules and alarms (4232581)

This article explains how the “PI Experiencing Errors” alarms work in Foglight for database cartridges (SQL Server, Oracle, SSAS, Azure SQL, etc.). These alarms let you know when the Performance Investigator (PI) stops collecting data for a period of time.

The “PI Experiencing Errors” alarm is triggered when PI data collection stops because of data gap errors, usually due to OLAP (Online Analytical Processing) data not being available or written to the performance repository as expected.

Each monitored agent type uses a hidden Agent Status Property (ASP) to determine how long a delay in SQL PI data collection is tolerated before triggering the rule.

• SQL Server, Azure SQL, MySQL PI, and PostgreSQL PI: paecOlapAvailabilityDataGapMinutes

• Oracle: spiecOlapAvailabilityDataGapMinutes

• SSAS: ssasSpiEngOlapAvailabilityDataGapMinutes

If SQL PI data has not been written to the performance repository within the configured threshold (default: 30 minutes) the alarm fires. 

These properties often appear as -2147483648 in the Agent Status Properties UI, which represents an unset (null) value using Java’s minimum 4-byte integer.

Related Rules

Depending on the monitored technology, the following rules may trigger using the same logic:

• DBSS - SQL PI: SQL PI Experiencing Errors (SQL Server)

• DBO - SQL PI: SQL PI Experiencing Errors (Oracle)

• SSAS - PI Experiencing Errors (SSAS) {This depends on the SQL Server agent also functioning with PI}

Note: The "MySQL_PI - PI repository availability" and "Postgres_PI - PI repository availability" rules are not based on these agent status property settings.   

Follow these steps to understand and address the alarm:

1. Data Gap Detection
   • The agent checks how long PI data hasn’t been collected. If the gap is longer than the allowed time the alarm triggers.
   • Short, occasional connection drops (where data resumes quickly) are normal and don’t require support.
2. Check Agent Status Properties
   • Look at the ASP settings for your agent. If the gap threshold property shows `-2147483648`, it’s unset and using the default.
3. Review OLAP Availability Settings
   • For each monitored type, the properties listed in the chart below control OLAP data checks:
4. Check Connections and Resources
   • Make sure the agent can reach the database and has enough memory, CPU, and disk space.
   • Fix network issues if gaps happen often.
5. Review Logs
   • Look in Foglight and agent logs for messages about OLAP or data collection errors.
6. No Action Needed for Brief Alarms
   • If the alarm clears on its own and PI data resumes, you do not need to contact support.
7. Contact Support if Needed
   • If alarms are frequent, long-lasting, or PI data never resumes, check Quest documentation or contact Quest support.

Property Table:

Note: A value of -2147483648 indicates the property is not explicitly set and defaults to the internal value (usually 30 minutes for data gap threshold).