Influx is a NoSQL, non-relational database, written from scratch, that optimizes for properties like higher compression and column scans for specific label sets. While it supports something SQL-ish, the query language is fairly limited, and I think they are moving further away to a more declarative, non-SQL language.

On the other hand, TimescaleDB has figured out how to re-engineer PostgreSQL to make it scale for time-series data. In doing so, it:

- Supports full SQL, including complex WHERE predicates, secondary indexes, non-time based aggregates, sub-queries, window functions, CTEs, etc.

- Allows you to store regular relational tables alongside your time-series data and perform arbitrary JOINS.

- Supports both full schema-tized and JSON/JSONB data models.

- Inherits the full ecosystem of Postgres (visualization like Tableau or Grafana, ORMs like Ruby and Django, connectors like Kafka, etc.). If it speaks to Postgres, it speaks to TimescaleDB.

- Also works with other Postgres extensions, including things like postgis for geo-spatial analysis.

- Enjoys the reliability of Postgres, including streaming replication for HA, clustered read replicas, point-in-time recovery, incremental backups, etc.

For more details: http://docs.timescale.com/