Open The-Riddles opened 7 years ago
Hi @The-Riddles: We do not currently support this feature, although it has come up a few times and have it on our roadmap.
I'm curious if there are functional or operational reasons driving this preference/requirement? From a use perspective, your SQL query to the hypertable for the two should be identical (and the performance differences should be minimal).
[As always, if other users have express similar preferences, please do upvote the request. Thanks!]
Our team use timescaleDB to store business monitoring data.
When major business to be used searching for display 00:00:00 to 00:00:00 which is general time ranges data.
(Typical is when we ask some questions like ‘Which cell has the highest 234g traffic today?’, time boundary ranges of today defined with start 00:00:00)
Like this, we use tsdb to searching data will crossing 2 chunks to find our business data when time ranges begin 8:00:00 to 8:00:00 of next chunk boundary.
We want search business sence data in 1 chunk so we split hypertable chunk with 1 day,
But default time ranges of 1 day chunk between 8:00:00 to 8:00:00 is not goodidear.
We want use chunk ranges begin 00:00:00 to 00:00:00 of next boundary that we can search any business data in 1 chunk every time.
So i think this is general requirements to custom datetime(ex: day1 10:00:00 ~ day2 10:00:00 or day1 14:00:00 ~ day2 14:00:00) split hypertable chunks boundary when depend on business data like questions above.