We are going to make small adjustments on the infrastructure of our public endpoint. The downtime should be minimal, if any, but spurious RPC failures can be expected during the maintenance window.