Cleansing the Pulse data

Pulse constantly monitors your TM1 instances and stores the information in a h2 database (or SQL Server if configured) and an Elastic cluster (Pulse v6 or later).

• h2 (or SQL server): It stores all the information that Pulse is monitoring plus Pulse configuration.
• Elastic cluster: It stores all the data second by second.

To know the status and the size of these databases, go to the Pulse Status page on the About page. In the example below the h2 database size is 295 MB and the Elastic cluster is 24,381 MB:

The Pulse.cfg file includes a Maintenance section to help you manage the growth of the databases.

By default, the maintenance is deactivated, Pulse will keep all data from the first time it was installed in its database.

To set a maximum size of the Pulse for TM1 folder, you can update the MaximumPulseDiskSpaceGB parameter from the Pulse.cfg file. If set to 20, once the Pulse folder reaches 20 GB, Pulse will start deleting the oldest data until the size comes back under 20 GB.

To reduce the size of the database, you can now easily choose to keep only the last 360 days of history.

To do that, a Maintenance section has been added to the conf/Pulse.cfg file including 5 parameters:

• MaintenanceJobHourInterval: The maintenance will execute by default each Sunday at 4:30 am (default: 0 30 4 ? * SUN *). To change this time, you will need to add a new CRON expression, you can use a CRON generator (http://www.cronmaker.com/)
• ServerHistoricDays: The number of days of server history data to keep (Default: 0, means non-active)
• ProcessHistoryHistoricDays (Pulse 5 only): The number of days of process history data to keep (Default: 0, means non-active)
• DumpFilesHistoricDays (Pulse 5 only): The number of days of dump files data to keep (Default: 0, means non-active)
• PulseLogFilesHistoricDays: The number of days of Pulse logs data to keep (Default: 0, means non-active)

If you are not familiar with the Pulse Elasticsearch data, please read the following article first:

• Understanding Pulse Elasticsearch data

To reduce the size of the Elasticsearch cluster, you can choose how many years, months and days of data you want to keep per indices.

To do that, a new Maintenance section has been added to the conf/Pulse.cfg file including 3 parameters:

• ElasticHistoricYears: The number of yearly indices to be kept by Pulse.
• ElasticHistoricMonths: The number of monthly indices to be kept by Pulse.
• ElasticHistoricDays: The number of daily indices to be kept by Pulse.

More information on Pulse indices here.

To activate the maintenance job:

1. Stop Pulse services
2. Update the above Maintenance settings in the conf/Pulse.cfg
3. Start Pulse services