It provides automatic retention policy selection (solves issue #7198), automatic reduction of precision for large time intervals (in accordance with the
max-row-limit parameter in InfluxDB).
Additionally, it fixes issues with partial intervals for
SUM() ... GROUP BY time() queries (issues #8010 and #6451).
We had issues with people panicking when seing sudden drops in curves, that where in fact false positives. This has completely solved the issue.
We’ve been using it for 2 years now and decided to open source it as many should be facing similar use-cases.