Company
Date Published
Author
Ryan Betts
Word count
1665
Language
English
Hacker News points
2

Summary

InfluxDB is a time series database that stores data in points with four components: measurement, tagset, fieldset, and timestamp. Points are associated with a measurement, which can have different tagsets or fieldsets. The data model includes a series, which is a group of points sharing the same measurement, tag set, and field key. InfluxDB receives points from clients through an HTTP POST endpoint, where each point is stored in a database and retention policy. Points are written to a write-ahead log (WAL) for durability and an in-memory cache for queryability. The WAL and cache work together to make incoming data durable but are insufficient for long-term storage. InfluxDB organizes its columnar format into time-bounded chunks, which are stored on disk as TSM files that can be compacted over time. The retention policy configures the duration, replication factor, and shard duration of a database, allowing users to manage their data's lifespan and query performance.