-1

According to this article the only real benefit influx has over postgres (using time-based indices) is space used. PostgreSQL is more performant, with time indices.

Why would one use influx-db over postgresql, then?

I heard from someone that aggregate functions like avg/count are faster in influx, but couldn't find any results online backing this up.

1

benefit influx has over postgres (using time-based indices) is space used.

Why would one use influx-db over postgresql, then?

Because one will be storing a lot of data in that sort of format, so space could be a limiting (or cost incurring) issue. Note that space used by a live database in disk is not the only effect this will have:

  • The database being bigger may mean more IO for the same queries, so even if neither is faster than the other once data is in RAM there could be a significant difference for queries touching "cold" data.
  • A database that is bigger (for the same amount of data) is likely to need more RAM to function optimally which could affect your hosting costs noticeably in extreme cases.
  • A bigger database will take longer to backup, impacting your maintenance plans.
  • Also the resulting backups will be bigger, impacting your storage requirements again and possibly CPU cost if compressing the result to reduce space used.

are faster in influx, but couldn't find any results online backing this up.

If you want us to comment on specific claims, then you'll need to link to those claims, otherwise we can only give an "it depends" response because there could be many factors that will affect such assessments.

I would assume that such claims are at least based in some fact as InfluxDB is optimised for that use case where-as Postgres supports it as an add-on feature, but that is not at all guaranteed to be the case. The only way to be sure in the absence of trustworthy benchmarks is to setup databases in both engines and perform benchmark tests. I would recommend that you do this if time permits rather than relying upon arbitrary benchmark from elsewhere. Doing it yourself means that you will test with your own applications' data and access patterns thereby getting a measure of how each will work for your application specifically, instead of how well they work in a more artificial construct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.