How do Time Series databases (like Gorilla/Prometheus) achieve 10x compression on timestamp data?

1. A Delta-of-Delta Encoding, which stores only the difference between a timestamp and its predicted value based on a fixed, previously observed interval.
2. B Generic GZIP Streaming, which utilizes a massive, high-latency dictionary to compress blocks of thousand timestamps into a single, opaque binary blob.
3. C Precision Truncation, where the database engine automatically drops millisecond or nanosecond data until the resulting integers are small enough to pack.
4. D Binary Deduplication, which identifies and removes every duplicate timestamp entry from the influx stream to ensure only one unique point exists per second.