Hi all,
We have created some Edge audiences but we do not see they are being evaluated in real-time. These audiences only have one attribute each which is updated daily via S3 batch connector.
When analyzing the process for some users we see the following:
1. Data is ingested in the AEP Datalake via S3 connector at 10:00h-10:05 UTC.
2. Data is in Profile at 10:30-10:40 UTC.
3. Edge segments are evaluated at 17:47h UTC. This time is the same at which both Streaming and Batch segments are evaluated.
We believed AEP evaluates Edge audience membership in real-time or near-real-time as data is ingested, and that as soon as a user attribute meet the criteria for an audience, they are added to or removed from that audience at the edge. However, we see Edge evaluation occurs as batch segmentation.
This process is crutial since we want to personalize the website via Adobe Target with fresh data.
Is here something we are misunderstanding?
Thanks!
해결되었습니다! 솔루션으로 이동.
토픽은 커뮤니티 콘텐츠를 분류하여 관련성 있는 콘텐츠를 찾는 데 도움이 됩니다.
조회 수
답글
좋아요 수
We need to consider the data processing of components and timing. The profile attribute can be evaluated in the edge segmentation, but the following is a process on how it must flow.
Data ingestion batch > data directed to datalake > data prompted to profile > segmentation evaluation > profile projection to edge> edge evaluation.
Hi @DavidRoss91 ,
Is there a way of changing this? According to the documentation, any segment definition referring to only a profile attribute could be evaluated as Edge:
@jlasso71 Edge segmentation can be validated using the edge API as noted in the documentation below.
https://developer.adobe.com/data-collection-apis/docs/endpoints/interact/
A profile only attribute is edge eligible for segment evaluation.
https://experienceleague.adobe.com/en/docs/experience-platform/segmentation/api/edge-segmentation
As you have indicated in your post above.
data was ingested 10:00h-10:05 UTC.
Data is in Profile at 10:30-10:40 UTC.
@jlasso71 when does the segment evaluation job complete? This will promote the data available for activation to the edge data stream.
We need to consider the data processing of components and timing. The profile attribute can be evaluated in the edge segmentation, but the following is a process on how it must flow.
Data ingestion batch > data directed to datalake > data prompted to profile > segmentation evaluation > profile projection to edge> edge evaluation.
Hi @brekrut ,
Segment evaluation job was completed at 17:47h UTC, but we have updated the schedule so that it starts at 11:00h UTC now (once Profiles are updated with fresh data coming from batch connector). By doing this, the segments have most recent data only one hour after ingestion.
I understand that if a profile attribute is updated via Streaming HTTP API, we would be able to evaluate Edge segmentation in real time for this case?
Thanks!
@jlasso71 Did you find the suggestions helpful? Please let us know if you require more information. Otherwise, please mark the answer as correct for posterity. If you've discovered a solution yourself, we would appreciate it if you could share it with the community. Thank you!
조회 수
답글
좋아요 수