Ongoing indexing

For an overview of the Algolia B2C jobs, see Indexing Overview.

You can update your Algolia index with data from different B2C objects on an ongoing basis as follows:

• product data:

  • AlgoliaProductIndex_v2 job (full catalog indexing or reindexing)
  • AlgoliaProductDeltaIndex_v2 job (ongoing indexing)

• price data:

  • AlgoliaProductPriceIndex_v2 job

• inventory data:

  • AlgoliaProductInventoryIndex_v2 job

• category data:

  • AlgoliaCategoryIndex_v2 job

You can schedule these jobs to run with different frequencies according to your business needs.

The product indexing jobs are configured for single-instance indexing out of the box. This means that in a live environment, they should all be run from the same Production instance. To learn how to configure the jobs so that multiple instances update the same index, see Mult-instance indexing.

The AlgoliaProductDeltaIndex_v2, AlgoliaProductPriceIndex_v2 and AlgoliaProductInventoryIndex_v2 jobs form a system. Together they export your product, pricing, and inventory information, respectively.

For more information about indexing your entire catalog, see Full catalog indexing.

Product data index jobs

AlgoliaProductDeltaIndex_v2

This job performs delta updates on your Algolia product records using the B2C Delta Export feature. The job retrieves the list of productIDs from the output file of the B2C delta export, returns the product objects from the database, enriches/transforms them and sends them to Algolia, fully replacing the product object with the same objectID in the index.

Which objects are included: B2C Product objects assigned to the site that have changed since the last job run (either via BM or a product feed import), enriched with price and inventory information. Changes to master products will cause its variants to be included in the delta as these can inherit attributes from the master.

Only changes to B2C Product attributes will cause a Product to be included in the delta update, changes to price and inventory information will not. However, once a product is included, it will have its price and inventory information indexed. Use the price and inventory jobs to index changes to prices and inventory, respectively.

The Salesforce B2C Delta Export mechanism doesn’t track changes of dynamic values, such as the online status of a product changing at specific times. That’s why you need to use the AlgoliaProductIndex_v2 job, or a create a custom job based on the algoliaProductIndex step to update changes in dynamically calculated values.

Record update method: fullRecordUpdate (default) - each exported product object fully replaces the existing Algolia product record with the same ID in the index. It includes changes to existing products, as well as the addition and removal of products. Fully replacing the product record in the index means that attributes that are not selected for indexing will be removed from the record in the index. For more information see Full record updates.

This also deletes old/stale properties from the Algolia index that were removed since the last export.

The record update method can be changed to partialRecordUpdate via the job’s indexingMethod parameter. This mode performs a partial update on the records in the index, meaning that only the attributes that are sent during the job run are updated for each record, leaving the rest intact. This mode can be used for multi-instance indexing, where not all ongoing jobs are run from the same Salesforce B2C instance. You can use the attributeListOverride job step parameter to define a smaller set of attributes to update when using partialRecordUpdate. For more info on partial record updates, see here.

Before you begin

• Create a support ticket with Salesforce to enable the Delta Export feature and its prerequisite, the Change Log feature on the desired instances.
• [optional] If you want to include changes that occur in production via replication, the Change Log Replication feature must also be active. All three feature switches must be activated in Staging and Production.

Job configuration

You can find the job configuration in the Job Steps tab.

Scope

Change the scope to Specific Sites and select all the sites you’d like to run the jobs for.

Job parameters

• catalogIDs (comma-separated values): add the catalogs you’d like to export the product changes from;
• consumer (default: algolia): change this if you’d like the B2C delta export job to use another work directory (see below);
• deltaExportJobName (default: productDeltaExport): the B2C catalogDeltaExport job step will automatically create a delta export job with this ID under Site Development > Delta Exports. This parameter, in conjunction with consumer, determines the work directory inside Impex (Impex/ src/platform/outbox/<consumer>/<deltaExportJobName>/).

algoliaProductDeltaIndex job step parameters

• consumer: controlled by the consumer job parameter, leave them linked
• deltaExportJobName: controlled by the deltaExportJobName job parameter, leave them linked
• attributeListOverride (comma-separated values, empty by default): if set, only this list of product attributes is exported by the job, otherwise the base attributes and the values from the Additional Product Attributes site preference are used.
• indexingMethod: fullRecordUpdate (default) or partialRecordUpdate, as for the preceding Record update method.
• failureThresholdPercentage (integer, default: 0): allowed percentage of records failing to be indexed before marking the job in error. Records successfully sent are kept in the Algolia indices.

For more information, see Indexing attributes.

Price and inventory data index jobs

The AlgoliaProductPriceIndex_v2 and AlgoliaProductInventoryIndex_v2 jobs are highly optimized, targeted jobs to update price and inventory data, respectively.

Each of these jobs iterates over all B2C Products assigned to the configured site and updates price/inventory information for your Algolia products via a partial update.

Which objects are updated: each B2C Product object assigned to the site, but the exported data only contains the product’s ID and

• (by default) the price Algolia product property for AlgoliaProductPriceIndex_v2 and
• (by default) the in_stock Algolia product property for AlgoliaProductInventoryIndex_v2.

For more information about extending the list of attributes, see Indexing attributes.

Record update method: partial - the product objects this job sends to Algolia will not fully replace the existing Algolia product record in the index, but will only update (by default) the price/in_stock property of the object. If the record doesn’t exist, a new one will be created.

Due to the partial update nature of these jobs, they can only add/update price and inventory data for product objects in your Algolia index, but cannot remove any existing properties.

Performance considerations

The products to be sent are broken down into chunks (chunk size can be configured) and these chunks are processed and sent concurrently. The number of concurrent threads the job runs on is determined by the B2C platform itself and is dependent on the available resources (should be higher on PIG instances).

The job also uses a retry strategy similar to other API clients.

In the case of both jobs the list of properties to be exported can be extended, but this may increase the runtime of the job (see below).

Job configuration (Job steps tab)

Both jobs have the same configuration options. Any differences will be pointed out where applicable.

Scope

Change the scope to Specific Sites and select all the sites you’d like to run each job for.

algoliaProductPricePartialIndex and algoliaProductInventoryPartialIndex job step parameters, respectively

• attributeListOverride (comma-separated values): the list of Algolia properties that are exported by this job. Only the attributes defined here are exported by the job. Defaults are:
  • "price" for algoliaProductPricePartialIndex;
  • "in_stock" for algoliaProductInventoryPartialIndex.
• failureThresholdPercentage (integer, default: 0): allowed percentage of failed updates before marking the job as an error. Successful updates are kept in the Algolia indices.

If you extend the list of exported attributes, make sure to add the same attributes to the Additional Product Attributes site preference in the Algolia BM module as well so that the product delta export job also includes those attributes and updates them during its run.

For more details on extending the list of attributes, see Indexing attributes.

Chunk size

By default, all product jobs process and send product records in batches of 500 (the maximum is 1,000). This is called the chunk size, with chunks being processed and sent in parallel, depending on the resources available. You can change the chunk size in steptypes.json, under "@type-id": "..." > "chunk-size".

Setting the chunk size too high can result in an api.jsJSONStringLength quota violation. A quota warning will be issued, but the platform won’t prevent execution of the job because the quota is not currently enforced.

Category data index jobs

AlgoliaCategoryIndex_v2

Use the AlgoliaCategoryIndex_v2 job to update your category index. You should run it every time your category structure changes.

When should you schedule your jobs to run?

To keep your Algolia indices in sync with your B2C Commerce data, you can run the following jobs manually or on a regular schedule.

As a general rule of thumb,

• run the AlgoliaProductDeltaIndex_v2 job every time you know your product data has been updated;
• run the AlgoliaProductPriceIndex_v2 job every time you know your price data has been updated, such as when a promotion or a new price book has been activated;
• run the AlgoliaProductInventoryIndex_v2 job every time you know your inventory data has been updated; you can run this job frequently to keep your inventory in sync;
• run the AlgoliaCategoryIndex_v2 job every time you know your categories have been updated.

If you want to trigger Algolia index updates following a product/price/inventory import, you can consider creating a job that adds both of these steps to the same sequential job flow. For more info see the B2C documentation on job flows. For more recommendations on improving indexing performance, see Performance considerations.