Reducing the impact on Datalake service

Complete these steps before turning on the Datalake switch for non-production WMS tenants. These steps address the issue with Datalake services running out of memory when there are large volumnes of data to publish. When the tenant database is refreshed or the tenant is turned on for the first time, additional memory is required.

  1. Complete the Purge Jobs For Deleted Rows scheduled job.

    See Purging job for deleted rows.

    The purge job purges data older then 30 days in the Deleted Rows table at all tenant facilities. Purging data reduces the number of records in the Deleted Rows table.

  2. Check the history of the job to find out if the job was executed successfully.
  3. Create a Jira for Cloud Ops and request an R&D to the memory size and total number of containers for Datalake service. Increased memory size and containers are required until the tenant data is completely published to Datalake.
  4. If you are planning to enable Datalake for multiple non-production tenants, then enable Datalake sequentially for each tenant. Ensure that your first tenant is syncrhonized before enabling the next tenant. Do not enable all tenants at once.