This message was deleted.
# generals
Slackbot
10/19/2023, 9:29 AMThis message was deleted.
p
Piotr Dołowy
10/19/2023, 10:06 AMqueries from sys.segments are independent from metadatadb{pgsql). I guess 20s for 165k is ok. I have 750k segments and query took 30-40s on average
g
Guillaume Leclanche
10/19/2023, 10:12 AMOh I was thinking by mistake that the sys.* queries were kind of internal druid mapping of the metadata db
Guillaume Leclanche
10/19/2023, 10:13 AMIs your druid console significantly slow at displaying any value then ? (like over 10s of seconds)
p
Piotr Dołowy
10/19/2023, 10:26 AMno, queries are slow but console is working smoothly (for druid 0.22.1 and 27.0.0)
Piotr Dołowy
10/19/2023, 10:31 AMbut 'datasources' and 'segments' tabs are very slow. Especially managing load rules via web conlos is a nightmare with so much segments
g
Guillaume Leclanche
10/19/2023, 10:48 AMYeah I agree the UI part of the console is ok, it's the display of data that is terrible. I'm not convinced that the behaviour we're both seeing is alright. It is likely to have other impacts on the way the cluster itself maintains its data
Guillaume Leclanche
10/20/2023, 10:51 AMfwiw, I've increase (doubled) the memory of the coordinators and it's now around 10s in my case.
Guillaume Leclanche
10/20/2023, 10:52 AMnot a incredibly huge improvement, but it's still better