How to automatically recompress a compressed chunk on delayed data inserts (TimescaleDB 2.23.1)?

1

Hi team,

I’m using TimescaleDB 2.23.1 with compression enabled. I read the blog post about allowing DML operations in highly compressed time-series data, which explains how TimescaleDB handles inserts into compressed chunks by staging the data and periodically merging/recompressing automatically via background jobs. Here is link of blog post

My use-case:

  • I have a compression policy that compresses old chunks.

  • Occasionally, delayed data arrives for a timestamp range that is already compressed.

  • I want this delayed data to be inserted and then the existing compressed chunk to be recompressed automatically, without manually decompressing and recompressing the chunk.

My questions:

  1. Is there a built-in or recommended way in TimescaleDB to automatically trigger recompression of a chunk after delayed inserts into an already compressed chunk?

  2. Does TimescaleDB’s internal job scheduler automatically handle the recompression of chunks once new data is added to them, or do I need to run recompress_chunk() or adjust the compression policy?

  3. If manual actions are required (like recompress_chunk()), what is the best pattern to handle this for delayed inserts in production?

Thanks!