How does indexing internally work?

Hi all,

I am writing an article on Time Series storage and analysis, and I am now really focusing on Influx, as you can tell reading all my posts here.

I am deeply studying Influx internal architecture. In particular, right now I am wondering about indexing: I know that tags are automatically indexed, and that they are the only indexable part of the records. But, what happens when I insert a point? Do its tags get indexed immediately? Or indexes are periodically rebuilt on new data?


There’s a pretty comprehensive two-part blog on the InfluxDB internals here that might be helpful to you.

There’s also very good documentation on the indexing here.

We’re happy to answer more questions as they come up! Let me know if your questions isn’t answered by these sources. :slight_smile: