Hi all,
What’s the aim of the timezone option creating an add_columnstore_policy?
E.g. I’d like to create a policy that refers to UTC time. So it shall ignore DST changes and next_start shall move with DST changes. I would have expected that for such use cases the option timezone is available. But please see the following example:
CALL add_columnstore_policy('mytesttable', after => '390 days'::interval, schedule_interval => '6 months', initial_start => '2026-01-25 15:15:00+00', timezone => 'UTC');
SELECT * FROM timescaledb_information.jobs WHERE job_id >= 1000;
-[ RECORD 1 ]-----+---------------------------------------------------
job_id | 1020
application_name | Columnstore Policy [1020]
schedule_interval | 6 mons
max_runtime | 00:00:00
max_retries | -1
retry_period | 01:00:00
proc_schema | _timescaledb_functions
proc_name | policy_compression
owner | postgres
scheduled | t
fixed_schedule | t
config | {"hypertable_id": 1, "compress_after": "390 days"}
next_start | 2026-07-25 16:15:00+02
initial_start | 2026-01-25 16:15:00+01
hypertable_schema | public
hypertable_name | mytesttable
check_schema | _timescaledb_functions
check_name | policy_compression_check
I would have expected that next_start is at 2026-07-25 17:15:00+02, so still at 15:15:00 UTC time.
(I can reach my goal by defining the schedule_interval not in days/weeks/months but in hours.)
So it doesn’t seem the timezone option has an influence on that. In which situations the timezone option is relevant then?
Best regards,
Paul