[07:47:56] hello folks [07:48:02] very nice new website for kserve: https://kserve.github.io/website/0.7/modelserving/control_plane/ [09:03:48] left a note in the kserve slack channel about our doubts for tornado/asyncio/etc.. [09:03:56] let's see if a merciful soul answers :D [10:04:21] 10Machine-Learning-Team, 10observability: Improve ORES observability - https://phabricator.wikimedia.org/T299137 (10elukey) @Halfak thanks a lot for the detailed summary, it helps a lot! I would definitely like to have some logs indicating what happened during a certain timeframe, to be able to correlate spike... [11:25:02] 10Lift-Wing, 10Machine-Learning-Team (Active Tasks): Load test the Lift Wing cluster - https://phabricator.wikimedia.org/T296173 (10elukey) There was a constant throttling (cpu) for the queue-proxy containers, now it should be gone. The following link can be good to see how cpu/memory wise are doing with revs... [11:27:03] 10Lift-Wing, 10Machine-Learning-Team (Active Tasks): Load test the Lift Wing cluster - https://phabricator.wikimedia.org/T296173 (10elukey) I also left some questions in Kserve's upstream slack channel about how tornato/asyncio/etc.. are working and if we should follow specific guidelines to make our code as p... [11:28:59] * elukey lunch! [16:16:58] the kserve upstream folks gave me this link https://kserve.github.io/website/master/modelserving/v1beta1/custom/custom_model/ [16:17:08] so the TL;DR seems to be: [16:17:43] - the async workers are used only if we mark code as co-routine etc.. and it seems mostly helping in the transformer, when calling external http services etcc. [16:18:09] - the tornado workers have an async loop, and code runs in it [16:18:27] so blocking/cpu-bound code reduces parallelism a lot [16:18:53] - we could deploy ray workers (not still super sure how) in order to have a separate pool for models [16:19:02] so the tornado loop would only dispatch calls to the models [16:20:20] it may be as simple as adding the annotation to the code (last famous words) [17:33:08] 10Lift-Wing, 10Machine-Learning-Team (Active Tasks): Load test the Lift Wing cluster - https://phabricator.wikimedia.org/T296173 (10elukey) From IRC: ` the kserve upstream folks gave me this link https://kserve.github.io/website/master/modelserving/v1beta1/custom/custom_model/ so the TL;DR seems to be: - the...