vince v1.1.7

Faster and more efficient queries by doing sequential scans. We carefully layout keys in such a away that queries process data in sequential order.

This means data in the same resolution will be processed together, ensuring key/value prefix seeks only loads blocks with interesting data resulting in fewer data reads and faster response times.

Sequential scans allows us to forget about concurrency within a request bounds. Given that we target single vCPU boxes, we get smooth utilization of cpu between ingestion and queries.