During the latest eqiad swift / ms-be rebalances I've noticed the new hosts experience higher latency than the rest, especially around PUT/DELETE.
Some things off the top of my head that are worth investigating:
- perf top shows native_queued_spin_lock_slowpath and that made me realize we're not load-balancing IRQs across CPUs, but we should (similar to cp / lvs hosts)
- test rebalances with less weight (i.e. moving less partitions around)