Data Retention Policies
By default, data (facts) in TIBCO Service Performance Manager is persisted to a database. Unless deleted, the FACT table in the database grows with time. TIBCO Service Performance Manager enables you to purge the older data from the system. Facts are deleted based on the created_time of the fact.
Retention Policy for FACTS
The older data can be purged periodically from the system using the following settings defined in the schema XML file:
<schema name="AMX_3_0" display-name="ActiveMatrix Service Grid 3.x"> <retention-policies> <retention-policy type="fact" period="1" unit="DAY" purge-time-of-day="-1" purge-frequency-period="3600000"/>
Where:
- type: Set the value of this attribute to "fact" to configure a retention policy for facts.
- period: A number, to be used along with the attribute "unit". These two attributes together define how much data is to be retained.
For example, setting period=1 and unit="DAY": If you set period=1 and unit="DAY", all facts older than one day are purged. If you set period=2 and unit="WEEK", all facts older than two weeks are purged. Possible values for unit are "DAY", "WEEK", and "HOUR".
- purge-time-of-day: Specify the time of the day when the purge activity would be scheduled. It takes the format HHMI where HH is the hour of the day and MI is the minute of the hour. For example, setting it to 0100 would schedule the first purge activity for 01 AM. Use -1 if you want the purge activity to be repeated every purge-frequency-period.
- purge-frequency-period: Specify how often to run the purge activity. Specify the time period in milliseconds.
Retention Policy for Metrics and Hierarchies
Use a retention policy in the schema XML to configure the retention policy for dimension hierarchies. For metric hierarchy tables, metrics are purged based on updated_time of the metric. The value of the type attribute should be a fully-qualified name of a dimension hierarchy as defined in the schema. All other attributes have the same meaning as defined in the previous section. For example, the first line in the following example, prompts TIBCO Service Performance Manager to purge all metrics data that is older than 2 weeks, for the hierarchy DevNodeCube/SrvTrends. It is configured to run at 01 AM and with a frequency of 1 day (86400000 milliseconds).
<retention-policy type="DevNodeCube/SrvTrends" period="2" unit="WEEK" purge-time-of-day="0100" purge-frequency-period="86400000"/>
Similarly, you can configure the retention policy of all other hierarchies by modifying the schema file.
The defaults for TIBCO Service Performance Manager are as shown in the following table:
Retention Policy Time | Period | Unit | Purge Time of Day | Purge Frequency Period |
---|---|---|---|---|
fact |
1 | DAY | -1 | 3600000 |
DevNodeCube/SrvTrends |
2 | WEEK | 0100 | 86400000 |
DevNodeCube/ApplTrends |
2 | WEEK | 0110 | 86400000 |
DevNodeCube/NodeTrends |
2 | WEEK | 0120 | 86400000 |
DevNodeCube/By Service |
2 | WEEK | 0130 | 86400000 |
DevNodeCube/SvcInstTrends |
2 | WEEK | 0140 | 86400000 |
DevNodeCube/GrpByClientIPRes |
2 | WEEK | 0150 | 86400000 |
DevNodeCube/GrpBySrvClientIP |
2 | WEEK | 0200 | 86400000 |
DevNodeCube/HttpResource |
2 | WEEK | 0210 | 86400000 |
DevNodeCube/JdbcResource |
2 | WEEK | 0220 | 86400000 |
InfCube/InferredStatus |
2 | WEEK | 0230 | 86400000 |
Assets/env |
2 | WEEK | 0240 | 86400000 |
Assets/app |
2 | WEEK | 250 | 86400000 |
Assets/node |
2 | WEEK | 300 | 86400000 |
Assets/node_hist |
2 | WEEK | 0310 | 86400000 |
Assets/svcinst |
2 | WEEK | 0320 | 86400000 |
Assets/svcinst_hist |
2 | WEEK | 0330 | 86400000 |
Assets/http |
2 | WEEK | 0340 | 86400000 |
Assets/http_hist |
2 | WEEK | 0350 | 86400000 |
Assets/jdbc |
2 | WEEK | 0400 | 86400000 |
Assets/jdbc_hist |
2 | WEEK | 0410 | 86400000 |