The PowerConnect Cloud agent can be configured to use checkpointing. Checkpointing is a technique for storing and retrieving the point at which an input should resume the next time it’s executed. The agent supports both ID and Timestamp based checkpoint values.

To avoid processing very large volumes of data if the PowerConnect Cloud Agent has been down for an extended period of time a Max Intervals to backfill setting is applied when reading the checkpoint value. This defaults to 15 so for example if an Input is set to run each minute the default maximum look back for a timestamp based checkpoint will be 15 minutes.

Note - currently in version 1.0.8 only the CPI message input and the CPQ event log input support checkpointing but others will be supported in the next version 1.0.9.

Configuring a Checkpoint for an Input

  • Whilst creating an Input or editing an existing Input click on the Checkpoint tab

  • Choose the Checkpoint Type

    • Timestamp - this uses a timestamp to store the last timestamp/date the Input read for the events it collected in its last run (defaults to current time on the first run)

    • ID - this uses a numerical ID field to store the last ID the Input read for the events it collected

  • Choose the Checkpoint Field

    • This is a list of known fields from Input Field Mapping that matches the Checkpoint Type e.g for Timestamp only the fields mapped to Date/Timestamp will be shown

  • Click Save

  • Under the Checkpoints Tab you can see the newly configured Checkpoint, its current value and the last time it was updated

Changing the Max Interval to Backfill setting

The default setting of 15 for the Max Interval to backfill can be changed

  • Click edit on the Input you wish to modify

  • Click the Advanced Tab

  • Change the value of the Maximum number of intervals to backfill

  • Click Save

    Note - Setting a very high value here could lead to high data lag and a high number of API calls to the target SAP API when the PowerConnect Cloud Agent needs to backfill large volumes of data