Monitoring gaps due to server reaching an overload state and data being discarded (4374403)

Monitoring gaps are present in multiple graphs for different agent types; the Foglight Management Server (FMS) logs have frequent messages for batches of data being discarded due to reaching an overload state and the Data Cache Eviction Monitor being active.

WARN  [QuartzScheduler.Monitor_Worker-5] com.quest.nitro.service.data.DataService - The server has reached an overload state. 26,066 batches of data have been discarded.

WARN [QuartzScheduler.Monitor_Worker-2] com.quest.nitro.service.monitor.AbstractMonitor - Data Cache Eviction Monitor is still active. Scheduled execution will be skipped.

Retention policy with a too small granularity (1ms) for the top_sqls_snapshots property of the SQL_Profiler_Model topology type can increase the amount of processing performed by the server.

WORKAROUND

Complete the following steps to alter the retention policy for the top_sqls_snapshots property of the SQL_Profiler_Model topology type:

1. Go to: Administration > Data > Manage Retention Policies.
2. Select SQLProfiler from the Filter By Cartridge drop-down list.
3. For the first retention policy of SQL_Profiler_Model > top_sql_snapshots, change the Age from 15minutes to 1minutes and the Roll-Up Period from 1ms to 0ms.

   Note: The roll-up period is set to zero for no aggregation to be performed.

   Before:

   

   After:

   

STATUS

This issue has been logged as defect Id. FOG-9241 and has been fixed in the 7.3.0 and higher releases of Foglight for Databases.