With T357392 proceeding, we 'll need to switch over some capacity from the parsoid cluster to the kubernetes cluster (we don't have unfortunately the spare capacity to move everything and then just reimage). This task is about tracking that work.
For eqiad, where the cluster is currently idling, I think i 'll be going in batches of 8 hosts at a time, aiming to be done way before the March Switchover.
For codfw, things are more difficult. Based on my grafana cluster overview reading, a good batch size appears to be 5
Final status
Of all the parse* hosts, parse1001 and parse1002 haven't been reimaged and will be reimaged in a followup task as part of parsoid-php service cleanups.