Change Details

Since the significant-events endpoint can be flaky, let's consider creating a separately hosted file that returns an either on off value for the app to consider before requesting the endpoint, or a scaled ratio for the app to consider before putting someone in an A/B testing bucket. The latter would give us the ability to scale back both the A/B test sampling determination and give us a form of traffic control at the same time.

We need to update the announcements endpoint to have the new article list for the experiment round of the exposure survey. We also want to add a new field that will contain a 0 - 100 percentage value to represent which percentage of users would fall into the experiment bucket. This will give us the ability to configure this value server-side, should we want to traffic control things based on survey response rate and/or the server load. We should also increase the version number of the survey announcement both server-side and client-side.

Since the significant-events endpoint can be flaky,We need to update the announcements endpoint to have the new article list for the experiment round of the exposure survey. We also want to add a new field that will contain a 0 - 100 percentage value to represent which percentage of users would fall into the experiment bucket. let's consider creating a separately hosted file that returns an either on off value for the app to consider before requesting the endpointThis will give us the ability to configure this value server-side, or a scaled ratio for the app to consider before putting someone in an A/B testing bucket. The latter would give us the ability to scale back both the A/B test sampling determination and give us a form of traffic control at the same timshould we want to traffic control things based on survey response rate and/or the server load. We should also increase the version number of the survey announcement both server-side and client-side.