Influx is a time series database sporting a SQL style interface and query language.
Instead of tables we use measurements. Every measurement contains tag sets (mandatory, indexed) the tuple of which forms a series. I.e, INSERT info,host=bob,location=eu and INSERT info,host=bob,location=uk forms two series (bob,eu) (bob,uk). Every measurement also contains field sets, the (potentially NULL / empty) data for that measurement. All fields have to have time.
Select a database
USE node_health;
Insert a measurement with one field
INSERT machine_state,host=b1,location=uk cpu_use=7.5;
Select a specific host from a measurement
SELECT * FROM machine_state WHERE host = 'b1';
Select mean CPU usage for a given host
SELECT mean("cpu_use") FROM machine_state WHERE host='b1';
Only the last 5 minutes
SELECT mean("cpu_use") FROM machine_state WHERE host='b1' GROUP BY time(5m) LIMIT 1;
Retention policies help data stop piling up. Data is clustered into shards (Logical blocks of data clustered based on time). Data retention policies apply to shards, not inserts. Once a shard falls outside of its retention policy, all data in it goes bye bye.