-
タイトル
Monitoring gaps due to server reaching an overload state and data being discarded -
説明
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.
-
原因
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.
<life-cycle-impl id="16459a82-3817-416e-aa07-20958471d276" name="lifeCycleForSQL_Profiler_Model-top_sqls_snapshots">
<periods>
<life-cycle-period-impl id="8ae29a168edef91f018edefea3444ed7" age="86400000" granularity="9223372036854775807"/>
<life-cycle-period-impl id="8ae29a168edef91f018edefea3444ed6" age="900000" granularity="1"/>
</periods>
</life-cycle-impl> -
対策
WORKAROUND
Complete the following steps to alter the retention policy for the top_sqls_snapshots property of the SQL_Profiler_Model topology type:
- Go to: Administration > Data > Manage Retention Policies.
- Select SQLProfiler from the Filter By Cartridge drop-down list.
- 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.
-
Defect ID
FOG-9241
