Scaling Timeseries Data With YugabyteDB: Efficient Rollup And Purge
Rollup past data: keep summary, aggregate old events by day with YugabyteDB YSQL, reduce database size & improve visualization
When you load timeseries from sensors, you store all details for fine grained analysis. But keeping this level of detail, the database will grow. With YugabyteDB, this is not a technical problem because you can scale-out the cluster. But you probably don't need to keep this level of detail forever. One solution in NoSQL databases is to enable TTL (Time To Live) to purge old data when expired. You can do that with YugabyteDB YCQL tables (the Cassandra-like API). But with YSQL you can do better: keep a summary of old data, aggregated to a higher level, for example per day instead of per millise...