This was a bit of maintenance we performed due to some imbalance in the machines that handle our write load. For redundancy and availability we split this responsibility across many machines. Some recent backend instability had caused some of our machines to be handling an imbalanced amount of write load. This maintenance "rebalanced" the write responsibility and got things back where we like it.

This is a form of maintenance that is built in to Keen's infrastructure so the only impact was added latency for new events being available for querying.