In past articles I discussed the Batch Level Of Service capability and one of the most important questions I am asked is how to best implement this for on-premise and cloud implementations. Here are some best advice to use this capability quickly with the minimum of work:
- Target Tolerances. Create algorithm entries using the provided base algorithm types setting your warning tolerance and error tolerance for each target value and method of evaluation. These represent the target values for warnings and error tolerances based around metrics like elapsed time, error rate, throughput etc. Remember these algorithm types are reusable across many batch controls.
- The base algorithms support a warning and an error tolerance. They also supporting using the latest completed or in progress execution. The latter is useful for live tracking during the execution.
- Link Targets to Batch Controls. Add the appropriate algorithm (with the appropriate target value) as the Batch Level Of Service algorithm. Multiple algorithms are also supported for the algorithm to cater for multiple criteria. To optimize the effort in managing the configuration of Batch Control's it is recommended to:
- Configure the most important or critical jobs to your business first. This is an important subset to focus upon.
- Configure the other processes as needed. Not all batch controls need targets.
The Batch Level Of Service will now be assessed against the latest execution (depending on the configuration of the algorithm itself) and the target values returning the relevant status.
Once you configure those targets you can use the Health Check Screen (or REST API for integration to other monitoring tools) and other capabilities to determine the batch level of service. For example:
Note: In the Oracle Utilities Cloud Services you can set the level of service on the schedule itself to check batch windows.
For more information about the Batch Level Of Service, refer to the following related articles: