Running Historical Loads Efficiently
How to run historical loads efficiently.
The historical load process is a batch process that will run in the background while the Analytics engine continues to process and compute current data in real time. The historical load process is resource intensive and requesting the load to run for a large number of metrics and/or for long periods of time requires a significant amount of time to complete. The following are guidelines for running a historical load:
- Avoid using the “Load All” option to perform a historical load of all metrics. This should not be required except in rare instances. This process may require 12 to 24 hours to complete depending on the defined date range.
- Only run historical loads for metrics that are new or have been edited. Edited means that the mapping category related to the metric(s) has been modified or changed..
- Run historical loads for smaller periods of time. Break the runs into 3-month windows while still running them for only those metrics that have changed. Allow time after a historical load for the Time Series synchronization process to run. Do not expect the time series data to be updated with the completion of the historical load.
- If you have modified metrics related to Scheduling and you want those changes to be reflected for future schedules that have already been created or generated, you can run a historical load for a future time period.
- Do not run historical loads as a means to populate time series data. The load will only delay the synchronization process. The Time Series synchronization process runs every 8 hours.
