Description
If I set a continuous aggregate policy with a start_offset NULL, will it keep re-aggregating all the buckets regularly even if no data was modified in the original hypertable? Or, is there an optimization to skip updating time buckets that have had no data modified?
Why ask: because we occasionally have data arriving with dates being months old. It would be annoying if the continuous aggregate does not automatically pick them up eventually.
Follow up: is it possible to enable compression on such a continuous aggregate? The documentation says that "the compression policy needs the compress_after parameter greater than the start_offset parameter of the continuous aggregate policy", so my understanding is that this is not supported at this time. However, given that standard compressed hypertables support data ingestion, I wonder if this limitation is planned to be removed from compressed continuous aggregates as well.