Try TimeScaleDB extension to PostgreSQL

[aufdenkampe]

@emiliom suggested that we might explore the use of the new open-source TimeScaleDB extension to PostgreSQL, to massively improve the performance of streaming time series data?

See:

• https://www.timescale.com
• https://github.com/timescale/timescaledb/

[aufdenkampe]

@miguelcleon, we're still actively exploring this as an option, especially over InfluxDB, if we still have performance needs that we can't satisfy with the ideas in #7 and/or #6.

[horsburgh]

It would be good to get some feedback from @miguelcleon on this based on his most recent experience. He's been looking at TimeScaleDB over the past little bit and his initial reports were that he hadn't seen a big speedup, but he may have sorted some things out since then.

@aufdenkampe - I'm curious about your comment of exploring TimeScaleDB "especially over InfluxDB". We've seen some really amazing performance increases for a lot of us cases through using InfluxDB. I think TimeScaleDB is definitely worth looking at because it would be more integrated with PostgreSQL, but we've had some great success speeding things up using InfluxDB.

[miguelcleon]

@horsburgh @aufdenkampe I'm building a new timescale db right now for the Luquillo CZO database, with some different settings from my initial attempt. It is taking a long time to build the timeseriesresultvalues hypertable, which maybe fine (16.3 million records). I'll post again when it is done.

[emiliom]

FYI, Don has been using TimeScaleDB in his new gig and was very impressed by it -- at least last Fall.

I'd be happy to set up a conference call with him and y'all, if you think that'd be worthwhile.

I'm deliberately not using his github handle here, to minimize noise on him or the expectation that he should reply to this issue; they keep him really busy!