PEGA0106: Conflicting queries in the PostgreSQL database system
The PEGA0106 alert is generated when multiple conflicting queries cause a blockage in the PostgreSQL database system.
Reason for the alert
The PegaAESRemote rules, which are available as hotfixes for Pega Platform™ 7.2.2 and later, include a Database Deadlock Detection Information agent. This agent queries the PostgreSQL database to find out if any queries in the database are blocked. The agent sends the blocked queries and the blocking queries to Pega Predictive Diagnostic Cloud™, which creates the PEGA0106 alert and a corresponding case. Blocked queries are a symptom of performance and, potentially, stability issues.
For more information, see Cases in Pega Predictive Diagnostic Cloud.
Example message text
Blocked queries detected due to concurrently running conflicting queries.
Blocked query: truncate table pegadata.pegaam_alert28.
Blocking query: SELECT "PC0"."msgid" AS "MsgID" , "PC0"."problemcorrelation" AS "ProblemCorrelation" , MIN("PC0"."pzinskey") AS "pySummaryText(1)" , COUNT(ASTERISK) AS "pySummaryCount(1)" FROM pegadata.pegaam_alert "PC0" WHERE ( "PC0"."pxcreatedatetime" > $1 AND "PC0"."pxcreatedatetime" <= $2 ) AND "PC0"."pxobjclass" = $3 AND ( "PC0".pzTenantId = $4 ) GROUP BY "PC0"."msgid" , "PC0"."problemcorrelation" ORDER BY 1 ASC, 2 ASC
Perform one of the following actions:
- On the Enterprise landing page, click the DB Current Activity tab to view the list of active queries. Make note of queries that have been running longer than 90 seconds. If any of the queries listed in the case are still running, contact the database administrator to review and potentially stop the blocking queries and stabilize the system operations.
- Use the Event Viewer to find long-running queries that occurred soon after you received the notification. Evaluate the queries and the associated events and cases to improve query performance and avoid excessive locking or lock escalation in the future. Work with a database administrator to determine how to reduce serialization and contention.