[06:57:09] 10Lift-Wing, 10Machine-Learning-Team, 10Patch-For-Review: Install Istio on ml-serve cluster - https://phabricator.wikimedia.org/T278192 (10QChris) >>! In T278192#7158132, @elukey wrote: > Added a request to https://www.mediawiki.org/wiki/Gerrit/New_repositories/Requests for the new repo! The `operations/deb... [07:20:17] https://www.manning.com/books/knative-in-action looks interesting :) [13:14:52] after so much time debugging the helm chart for knative, I discovered that the issue was a wrong setting [13:15:15] basically [13:15:18] queueSidecarImage: {{ .Values.docker.registry }}/{{ .Values.core.queue_proxy.image }}:{{ .Values.core.queue_proxy.version }} [13:15:30] was the right one, meanwhile I was using .Values.core.controller.image [13:15:42] so the sidecar for queue-proxy was the controller [13:15:48] that of course doesn't work [13:16:20] aside from my sadness for this error, I found a very nice debugging trick with helm 3 [13:16:53] 'helm template' takes a chart and generates a yaml file [13:17:06] like serving-core.yaml that knative upstream offers [13:17:20] so I put the upstream one and mine side to side, and I quickly spotted the issue [14:49:31] oh damn, alright. Good to know. [15:14:35] elukey: niiiice! glad to hear it was just an incorrect [15:14:56] setting [15:20:26] hi! [15:20:43] I am still trying to figure out one little thing that doesn't work, but shouldn't be knative-serving related [15:20:52] the helm chart is in code review [15:21:15] I think that I have removed the rbac-proxy thing from kfserving and now it complains [15:21:24] but I'll try to figure it out asap [15:21:48] after that there is the kfserving helm chart, should be similar to the knative one but we'll have to add some TLS certs [15:22:45] chrisalbon: today's meetings are all cancelled right? (just to confirm) [15:24:31] ^ i think so [15:26:38] yeah all canceled [15:26:44] super [15:29:16] I'm mostly just watching all hands and writing emails [16:22:52] good news is that https://github.com/kubeflow/kfserving/commit/9995b0ebde1f53710745acc0bb5257aa140beb13 got merged [16:22:59] for the self-signed-ca script [16:33:58] \o/ [16:53:15] nice! [16:59:40] accraze: accraze/kfserving-custom-model works right? [17:00:06] I am seeing an issue with queue-proxy not able to contact 127.0.0.1:8080 for the readiness probe [17:00:23] so I was wondering if it was the kfserving-container or some other error [17:00:33] it is probably my fault again but better to double check [17:04:47] Hello everyone , This side anubhav . I actually needed help to get the data for my edit quality model . I got the rev_id's for each edit . But when I tried to use the wikimedia API to get the content w.r.t a rev_id , I got badrevid as the output [17:04:56] Please someone help me out with it . [17:11:18] anubhav_sharm: hi! The team is attending a virtual all hands during these days so we may have some delay in answering [17:11:42] can you post in here the call the wikimedia API that you are testing? [17:11:52] so we can more easily debug [17:25:11] * elukey afk for the evening [18:16:42] URL = "https://www.mediawiki.org/w/api.php" [18:16:43] PARAMS = { [18:16:43] "action": "query", [18:16:44] "prop": "revisions", [18:16:44] "revids" : 640469309, [18:16:45] "rvprop": "timestamp|user|size|comment|content", [18:16:45] "rvslots": "main", [18:16:46] "formatversion": "2", [18:16:46] "format": "json" [18:16:47] } [18:16:53] this is the code snippet that i am using [18:51:08] anubhav_sharm: it looks like the revision with id 640469309 does not exist on the wiki url you are using (mediawiki.org) [18:51:45] but it was there in the dataset that i generated from the editquality github repository . [18:52:27] like it was mentioned in the enwiki.human_labeled_revisions.20k_2015.json file [18:53:05] oh weird lemme take a look [18:57:46] ah ok i think i see what's happening [18:57:55] that revision exists on english wikipedia: https://en.wikipedia.org/w/index.php?oldid=640469309 [18:59:00] so in your code snippet, you should do `URL = "https://en.wikipedia.org/w/api.php"` instead [18:59:47] anubhav_sharm: ^does this make sense? [19:08:36] yeah makes sense accraze . Thank You :p [19:08:46] *ignore the emoji please