We are landing the following change to the Attribution Reporting API focused on: * Improving API report batching capabilities
This change is based on ad-tech feedback that we heard regarding the need for additional report batching flexibility so that different report contributions can be batched at different cadences. Currently aggregate reports generated by the API can consist of multiple separate contributions which are key:value pairs. API callers can batch together aggregate reports and then process them in the aggregation service, which consists of decrypting the reports, aggregating the values, and adding noise, before returning a summary report to the API caller. Additionally, currently all contributions in an aggregate report must be processed by aggregation service at the same time. With this change, the API will now allow API callers to specify filtering IDs as part of each contribution (i.e. each key:value pair) in the aggregate report. API callers can then use these filtering IDs, which will also be part of the encrypted payload of the report, to specify which contributions they want to have processed by the aggregation service at a given time. This will allow API callers to process contributions with different filtering IDs at different cadences.
Explainers: https://github.com/WICG/attribution-reporting-api/pull/1290