All trended data are stored in the Fathom Trend Database. Trending is enabled when Resource Monitoring is enabled to
Poll Resources and
Collect trend data:
Options > Resource Monitoring:
Poll resources
Generate alerts
Collect trend data
Include status changes
The size of the fathom database (by default found in
<fathominstalldir>/db/fathom.db) is dependant on:
A. The Number of events configured for monitoring.
- The more resources that are polled at a set interval the more memory is needed and this can lead to CPU spikes on the java process.
- Monitored resources record entries in the fathom trend database every time the scheduled polling cycle occurs and is set to trend. For example: if the schedule is configured to Poll every 5 minutes (300 seconds) and to Trend every Poll interval, then that information will be stored in the trend database every 5 minutes.
- The more resources monitored, the more data will be stored in the trend database.
- The more frequent the polling of resources and the lower the Trended value, the more data will be stored in the trend database.
1. Revise the current Poll interval and the Trend of monitored resources:
- Consider first if the default Polling interval of 300 sec is needed or can be extended
- Consider for each if the Trend needs to be recorded for every Poll interval or can be extended to every nth Poll instead.
2. Modify the Poll interval and the Trend of monitored resources:To change the Resource Monitor Defaults:Default
Poll and
Trend intervals are set when each resource is added as a monitored resource. Any changes at the Default level, will only take effect for
new resources added in future.
OPTIONS > User Preferences > Resource Monitor Defaults > Resource functions
< list of resources eg: "Database Resource Defaults", File Resource Defaults etc>To change Poll and Trend intervals for existing resources:Change the
Trend and Optionally
Poll for the resource itself.
This does not have to be done manually for every resource that needs changing, if they need changing to the same settings for each resource type.
1. First manually change these for one resource type:
Example: For a Database
Select a current monitored Database
> Monitoring Plans > Schedule Plan > EDIT : Advanced Settings :
Schedule: Default_Schedule
Polling Interval: 10 minutes
Alerts Enabled: true
Trend Performance Data: true
Trend values:
Trend Db_ActAPW every: 3 poll(s)
Trend Db_ActBuf every: 3 poll(s)
Trend Db_ActIOFile every: 3 poll(s)
Trend Db_ActIOType every: 3 poll(s)
Trend Db_ActIdx every: 3 poll(s)
Trend Db_ActLock every: 3 poll(s)
Trend Db_ActLog every: 3 poll(s)
Trend Db_ActRec every: 3 poll(s)
Trend Db_ActServ every: 3 poll(s)
Trend Db_ActSum every: 3 poll(s)
Trend Db_AreaStatus every: 3 poll(s)
Trend Db_Checkpoint every: 3 poll(s)
Note: If this managed resource doesn't need trending at all, like the fathom trend database itself, set:
"Trend Performance Data: false
2. Migrate this resource's monitoring to all of the same resources through:
OPTIONS > User Preferences > Distribute Resource Properties
- Selection Filters presented are for that resource type (eg the database) that has been modified. Only select "Remove and replace existing rules" if different/additional rules to that which has just been changed have not been configured on the target, or these really do need removing.
- Target Container can be the same container (eg several databases on the same container as the "source" or remote containers).
- Target Resources are the resources to copy these new monitoring rules to.
Resource Monitoring
Poll and
Trend rules can be applied with the resources running online, the fathom trend database and the fathom plugin running online. No need to shutdown.
B. How often the Data Compaction Job is run and with what Options.The size of the Fathom Trend Database can be managed by running the
Data Compaction Job.
Jobs > Job Templates > DataCompaction
A Job Template can be used to enable and run Data Compaction against the trend database at set intervals and associated Data Compact and Purge Options:
For detailed information, refer to OpenEdge Management: Trend database Guide and Reference
C. Tune the fathom database
The fathomtrend database is an OpenEdge database. Tuning it for the amount of monitored resources trend information that are collected is often overlooked.
For example,
1. The default Before-Image Cluster Size: 524288 can be increased to improve checkpoint times the next time this is offline:
$ proutil fathom -C truncate -bi 8192
2. The default Shared memory configured is only 4MB (1024x4), which can be improved without much effort as can for example the number of bi buffers
[configuration.fathomtrenddatabase.defaultconfiguration]
...
beforeimagebuffers=30 # For Enterprise databases, 60
blocksindatabasebuffers=32000 # this is the default for all OE 11 databases (125 MB), for enterprise databases 64000 (250 MB)
