Hello,
I am currently looking into TimeScaleDB to figure if we can use it in production. My experience with it is very limited.
My simple test setup looks like this. I have one (big) table that contains data like this:
timestamp, value a, value b, value c.
I created a hypertable using the timestamp and indexes for all values, since I want to filter on them. My main use case is aggregation with time bucketing.
When I ran subsequent queries, one query at a time, everything is fine and the DB is responsive. However when I do the same queries multithreaded, I noticed that it takes much longer (factor two to three) to complete all queries, vs doing it single threaded subsequently.
This is very surprising to me. Is this just how TimeScaleDB operates or am I doing something wrong? There are no foreign key relations and during querying there where no write operations. It is running in a single Postgresql instance. I queried it from a Java 8 program using HikariCP as connection pool.
Any advice would be highly appreciated. Thanks for reading. Cheers!
1 post - 1 participant