Conversation
Navy Federal Credit Union
US
Last activity: 11 Jul 2025 15:43 EDT
Elastic search using SRS
We are starting to use SRS on Pega cloud. Our existing apps use traditional elastic search. We are experiencing issues with embedded property search in SRS. I will explain further below.
Behavior:
We are searching using pxRetrieveSearchData activity by supplying pySearchString. This string is build dynamically like pySearchString = <embeddedPageList>.property:*<content>* e.g. pySearchString = MemberDetailsPageList.MemberAccessNumber:*00000000000022*
The expectation is that the search look for MemberDetailsPageList page list entries and specifically look for content (value contains) of property named "MemberAccessNumber". However it fails to do so. This works in the traditional elastic search.
Question:
Are there changes to elastic search indexing and retrieval of embedded properties in SRS? Any insights from the product team is appreciated as embedded properties search is very heavily used in many of our applications.
Pegasystems Inc.
US
There are differences in how SRS and embedded search work by default, especially with embedded properties. There are a few settings related to this you can change/create that may be necessary depending on the version of Pega in use (both in the Pega-SearchEngine ruleset):
- indexer/srs/includeCSPEmbeddedPropertiesInDataModel
- indexer/srs/indexAllFieldsForFTS
This is discussed a bit here: Compatibility of search engines upon migration
Setting indexAllFieldsForFTS to true is probably the easiest way to make the search/indexing more like it was with embedded search, but you will likely be indexing more properties than are needed. After changing either of the settings, you would need to update the data models on the Search Landing Page and then re-index any classes that needed this update.