Updating your Content Fragments for optimized GraphQL Filtering updating-content-fragments-for-optimized-graphql-filtering

To optimize the performance of your GraphQL filters, run a procedure to update your Content Fragments.

NOTE
After updating your Content Fragments, you can follow the recommendations for Optimizing GraphQL Queries.

Prerequisites prerequisites

There are prerequisites for this task:

  1. Ensure that you have a minimum of the 2023.1.0 release of AEM as a Cloud Service.

  2. Ensure that the user performing the task has the required permissions:

    • at a minimum, the Deployment Manager role in Cloud Manager is required.

Updating your Content Fragments updating-content-fragments

  1. Enable the update by setting the following variables for your instance using the Cloud Manager UI:

    Cloud Manager Environment Configuration

    The available variables are:

    table 0-row-8 1-row-8 2-row-8 3-row-8 4-row-8 5-row-8
    Name Value Default Value Service Applied Type Notes
    1 CF_MIGRATION_ENABLED 1 0 All Variable Enables(!=0) or disables(0) triggering of Content Fragment migration job.
    2 CF_MIGRATION_ENFORCE 1 0 All Variable Enforce (!=0) remigration of Content Fragments. Setting this flag to 0 does an incremental migration of CFs. This means, if the job is terminated for any reason, then the next run of the job starts migration from the point where it got terminated. The first migration is recommended for enforcement (value=1).
    3 CF_MIGRATION_BATCH 50 50 All Variable Size of the batch for saving the number of Content Fragments after migration. This is relevant to how many CFs are saved to the repository in one batch, and can be used to optimize the number of writes to the repository.
    4 CF_MIGRATION_LIMIT 1000 1000 All Variable Max number of Content Fragments to process at a time. See also notes for CF_MIGRATION_INTERVAL.
    5 CF_MIGRATION_INTERVAL 60 600 All Variable Interval (seconds) to process the remaining Content Fragments up until the next Limit. This interval is also considered as both a wait-time before starting the job, and a delay between processing of each subsequent CF_MIGRATION_LIMIT number of CFs. (*)
    note note
    NOTE
    (*)
    The value of CF_MIGRATION_INTERVAL can also help to approximate the total execution time of the migration job.
    For example:
    • Total number of Content Fragments = 20,000
    • CF_MIGRATION_LIMIT = 1000
    • CF_MIGRATION_INTERNAL = 60 (Sec)
    • Approximate time required to complete the migration = 60 + (20,000/1000 * 60) = 1260 Sec = 21 Minutes
      The additional “60” seconds added at the start is due to the initial delay when starting the job.
    This is only the minimum time required to complete the job, and does not include the I/O time. The actual time taken could be more than this estimation.
  2. Monitor the progress and completion of the update.

    To do this, monitor the logs on author and golden-publish from:

    • com.adobe.cq.dam.cfm.impl.upgrade.UpgradeJob

      • Author logs; for example:

        code language-shell
        23.01.2023 13:13:45.926 *INFO* [sling-threadpool-09cbdb47-4d99-4c4c-b6d5-781b635ee21b-(apache-sling-job-thread-pool)-1-Content Fragment Upgrade Job Queue Config(cfm/upgrader)] com.adobe.cq.dam.cfm.impl.upgrade.UpgradeJob This instance<dd9ffdc1-0c28-4d04-9a96-5d4d223e457e> is the leader, will schedule the upgrade schedule job.
        ...
        23.01.2023 13:13:45.941 *INFO* [sling-threadpool-09cbdb47-4d99-4c4c-b6d5-781b635ee21b-(apache-sling-job-thread-pool)-1-Content Fragment Upgrade Job Queue Config(cfm/upgrader)] com.adobe.cq.dam.cfm.impl.upgrade.UpgradeJob Scheduling content fragments upgrade from version 0 to 1, slingJobId: 2023/1/23/13/13/50e1a575-4cd7-497b-adf0-62cb5768eedb_0, enforce: true, limit: 1000, batch: 50, interval: 60s
        
        23.01.2023 13:20:40.960 *INFO* [sling-threadpool-09cbdb47-4d99-4c4c-b6d5-781b635ee21b-(apache-sling-job-thread-pool)-1-Content Fragment Upgrade Job Queue Config(cfm/upgrader)] com.adobe.cq.dam.cfm.impl.upgrade.UpgradeJob Finished content fragments upgrade in 6m, slingJobId: 2023/1/23/13/13/50e1a575-4cd7-497b-adf0-62cb5768eedb_0, status: MaintenanceJobStatus{jobState=SUCCEEDED, statusMessage='Upgrade to version '1' succeeded.', errors=[], successCount=3781, failedCount=0, skippedCount=0}
        
      • Golden-publish logs; for example:

        code language-shell
        23.01.2023 12:35:05.150 *INFO* [sling-threadpool-8abcc1bb-cdcb-46d4-8565-942ad8a73209-(apache-sling-job-thread-pool)-1-Content Fragment Upgrade Job Queue Config(cfm/upgrader)] com.adobe.cq.dam.cfm.impl.upgrade.UpgradeJob This instance<ad1b399e-77be-408e-bc3f-57097498fddb> is the leader, will schedule the upgrade schedule job.
        
        23.01.2023 12:35:05.161 *INFO* [sling-threadpool-8abcc1bb-cdcb-46d4-8565-942ad8a73209-(apache-sling-job-thread-pool)-1-Content Fragment Upgrade Job Queue Config(cfm/upgrader)] com.adobe.cq.dam.cfm.impl.upgrade.UpgradeJob Scheduling content fragments upgrade from version 0 to 1, slingJobId: 2023/1/23/12/34/ad1b399e-77be-408e-bc3f-57097498fddb_0, enforce: true, limit: 1000, batch: 50, interval: 60s
        ...
        23.01.2023 12:40:45.180 *INFO* [sling-threadpool-8abcc1bb-cdcb-46d4-8565-942ad8a73209-(apache-sling-job-thread-pool)-1-Content Fragment Upgrade Job Queue Config(cfm/upgrader)] com.adobe.cq.dam.cfm.impl.upgrade.UpgradeJob Finished content fragments upgrade in 5m, slingJobId: 2023/1/23/12/34/ad1b399e-77be-408e-bc3f-57097498fddb_0, status: MaintenanceJobStatus{jobState=SUCCEEDED, statusMessage='Upgrade to version '1' succeeded.', errors=[], successCount=3781, failedCount=0, skippedCount=0}
        

    Customers, who enabled access to the environment logs using Splunk, can use the example query below to monitor the upgrade process. For details about enabling Splunk logging, see Debugging Production and Stage.

    code language-splunk
    index=<indexName> sourcetype=aemerror aem_envId=<environmentId> msg="*com.adobe.cq.dam.cfm.impl.upgrade.UpgradeJob Finished*"
    (aem_tier=golden-publish OR aem_tier=author) | table _time aem_tier pod_name msg | sort -_time desc
    

    Where:

    • environmentId - a customer environment identifier; for example, e1234
    • indexName - a customer index name, gathering aemerror events

    Example output:

    table 0-row-4 1-row-4 2-row-4 layout-auto
    _time aem_tier pod_name msg
    2023-04-21 06:00:35.723 author cm-p1234-e1234-aem-author-76d6dc4b79-8lsb5 [sling-threadpool-bb5da4dd-6b05-4230-93ea-1d5cd242e24f-(apache-sling-job-thread-pool)-1-Content Fragment Upgrade Job Queue Config(cfm/upgrader)] com.adobe.cq.dam.cfm.impl.upgrade.UpgradeJob Finished content fragments upgrade in 391m, slingJobId: 2023/4/20/23/16/db7963df-e267-489b-b69a-5930b0dadb37_0, status: MaintenanceJobStatus{jobState=SUCCEEDED, statusMessage='Upgrade to version '1' succeeded.', errors=[], successCount=36756, failedCount=0, skippedCount=0}
    2023-04-21 06:05:48.207 golden-publish cm-p1234-e1234-aem-golden-publish-644487c9c5-lvkv2 [sling-threadpool-284b9a9a-8454-461e-9bdb-44866c6ddfb1-(apache-sling-job-thread-pool)-1-Content Fragment Upgrade Job Queue Config(cfm/upgrader)] com.adobe.cq.dam.cfm.impl.upgrade.UpgradeJob Finished content fragments upgrade in 211m, slingJobId: 2023/4/20/23/15/66c1690a-cdb7-4e66-bc52-90f33394ddfc_0, status: MaintenanceJobStatus{jobState=SUCCEEDED, statusMessage='Upgrade to version '1' succeeded.', errors=[], successCount=19557, failedCount=0, skippedCount=0}
  3. Disable the update procedure.

    note important
    IMPORTANT
    This step is required to complete the upgrade.

    After the update procedure has run, reset the cloud environment variable CF_MIGRATION_ENABLED to ‘0’, to trigger the recycling of all pods.

  4. table html-authored no-header
    table 0-row-8 1-row-8
    Name Value Default Value Service Applied Type Notes
    CF_MIGRATION_ENABLED 0 0 All Variable Disables(0) (or Enables(!=0)) triggering of Content Fragment migration job.
    note note
    NOTE
    This is important for the publish tier, as the content update is only done on golden-publish, and when recycling of pods, all normal publish pods are based on the golden-publish.

    Verify completion of the update procedure.

    You can verify the successful completion of the update using the repository browser in the Cloud Manager developer console to check the content fragment data.

    • Before the first complete migration, the cfGlobalVersion property does not exist.
      Therefore, the presence of this property, on the JCR node /content/dam with a value of 1, confirms the completion of the migration.

    • You can also check the following properties on the individual Content Fragments:

      • _strucVersion should have the value of 1
      • indexedData structure must exist
      note note
      NOTE
      The procedure updates Content Fragments on Author and Publish instances.
      Therefore, Adobe recommends that you perform the verification by way of the repository browser for at least one Author and one Publish instance.

Limitations limitations

Be aware of the following limitations:

  • Optimization of the performance of GraphQL filters is only possible after a complete update of all your Content Fragments (indicated by the presence of the cfGlobalVersion property for the JCR node /content/dam)

  • If Content Fragments are imported from a content package (using crx/de) after the update procedure is run, then those Content Fragments are not considered in the GraphQL query results until the update procedure is run again.

recommendation-more-help
fbcff2a9-b6fe-4574-b04a-21e75df764ab