Michael Sayapin

Perhaps it's worth noting that chunk size of `controller_log` is `1d`.

Sure, I can prepare a fuller example, but you can see from the listings above, when the view is created with `timescaledb_experimental.time_bucket_ng` explain shows `Filter: ((timescaledb_experimental.time_bucket_ng('1 day'::interval, created, 'Asia/Shanghai'::text) >...

```sql # CREATE TABLE test(t TIMESTAMPTZ, value INTEGER); CREATE TABLE # SELECT create_hypertable('test', 't', chunk_time_interval => INTERVAL '1h'); create_hypertable -------------------- (30,public,test,t) (1 row) # INSERT INTO test SELECT t, floor(sqrt(random()...

Even with an index on `timescaledb_experimental.time_bucket_ng('1 day'::interval, t, 'Asia/Shanghai'::text)` ng-based cagg is still scanning all the chunks @RafiaSabih

Any ideas? I think it's not cagg per se, but `time_bucket_ng` with TZ set is not excluding chunks from scanning. With `WITH NO DATA` cagg and `materialized_only => false` it...

@RafiaSabih any ideas? It's totally blocking us from using `time_bucket_ng`.

Totally useful! We currently use mosquitto this way. Cluster local apps connect over 1883 without authentication. Websocket listener is exposed via an Ingress and is authenticated.

A proposed fix would be changing `WORD = "[A-Z]*[a-z]*[0-9]*"` for `WORD = "[A-Z]*[a-z0-9]*"`.

Without `timezone` on `time_bucket` it works instantly.

On further digging, it seems like 2.8.0 `time_bucket` (with `timezone`) behaves exactly like pre-2.8 `time_bucket_ng` - doing a full hypertable scan from the continuous aggregate refresh code. If I set...