The heatmap features shows an event amplification behaviour that puts a high load on Kafka. As all experimental features, we should be able to turn it off to avoid incidents.
We can phase that option out when more granular config is implemented, by checking the team's PG config like blobby does.
Changes
Add a HEATMAPS_PROCESSING_ENABLED envvar, default to true
When set to false, strip the heatmap data from the event, but don't write to Kafka
Does this work well for both Cloud and self-hosted?
Problem
The heatmap features shows an event amplification behaviour that puts a high load on Kafka. As all experimental features, we should be able to turn it off to avoid incidents.
We can phase that option out when more granular config is implemented, by checking the team's PG config like blobby does.
Changes
HEATMAPS_PROCESSING_ENABLED
envvar, default to trueDoes this work well for both Cloud and self-hosted?
How did you test this code?
Unit tests