Currently we have storage disabled for all parsoid traffic on RESTBase except of enwiki.
On the previous deployments we saw an increase on cache prewarm jobs. The current assumption is that there is content on RESTbase that wasn't updated lately or doesn't get edits so there was a stored version of /page/html that wasn't available on ParserCache.

After a few hours of the deployments the amount of jobs and workers saturation was back to regular levels.
In order to handle the enwiki switchover we need to scale up the workers that are affected by parsoid traffic.