Continuous Queries Support

[Abdel K]

I am trying to evaluate the ability of KairosDB on performing continuous queries on time series data.

Continuous queries on time series data are queries that run automatically and periodically on real-time data (and maybe store query results in a specified measurement). Some Time Series Database Systems (TSDBs) have a built-in operation that allows doing this (e.g. InfluxDB, TimescaleDB).

I couldn't find in KairosDB' documentation a way to perform this type of query. Another alternative would be to program the frequency of queries in a separate programming language/system clock that would run the given query frequently and automatically every time length. This is however not guaranteed to be fair 1) to the system because the performance would depend on an independent factor (the third-party programming language) and 2) to the other systems because they provide an extra convenient feature that KairosDB does not.

My question is there an efficient way that could be used for the purpose of performing continuous queries? If not, what are the alternative I have using KairosDB in order to perform the continuous queries?

Thanks!

[Brian Hawkins]

We have a concept of rollups which is kind of what you are asking about.  The rollup periodically runs a query and saves the results as another metric.  This can be also done via the api by using the save_as aggregator on your query (same as the rollup does)

Brian