24 Jul 2024 |
Sheogorath | The kube-prometheus-stack chart utilises the operator but also deploys some useful defaults and provides some standard metrics. You can go without, but you'll probably reimplement parts of it. | 07:43:41 |
Peter Eriksson | That is not my question - my question is what others prefer. | 08:14:49 |
Rens Houben | And the answer to that would vary depending on 1) opinion 2) familiarity with either method and, as a consequence, 3) skill level with either. | 08:28:39 |
Rens Houben | The only general answer would be "Use whichever one you have an easier time working with" | 08:28:54 |
Peter Eriksson | Not looking for a general answer, I am asking what you personally prefer from your experience with both. | 08:29:38 |
Peter Eriksson | And if you are not familiar with both - you can’t answer. | 08:29:49 |
effendy | Beautiful. | 08:30:20 |
Rens Houben | Then I'm sorry to say you're probably not going to get many answers, and I hope you'll figure out what actual question you really should be asking. Good day. | 08:31:01 |
Sheogorath | In reply to @zkshnqvl:matrix.org That is not my question - my question is what others prefer. What I tried to say: the answer is both, a community chart that deploys and uses Prometheus operator. | 09:40:28 |
| Redstone changed their display name from redstone-menace to Redstone. | 10:16:36 |
| 0x3bb changed their profile picture. | 21:24:08 |
25 Jul 2024 |
| @haruska1234:matrix.org joined the room. | 06:20:42 |
| @haruska1234:matrix.org left the room. | 07:19:24 |
sudoer | https://forums.truenas.com/t/the-future-of-electric-eel-and-apps/5409
K3s getting pulled out in favor of docker-compose | 15:20:59 |
sudoer | My guess is the truenas helm charts go away. I never liked them, tried installing one on my standalone k8s cluster, I don't remember what specifically I objected to but I remember thinking it was very rude. Stepping on things I had already spent time setting up on my own. | 15:24:21 |
sid | I agree with their change with the argument that a NAS isn't the ideal place to run a bunch of other services (with some exceptions). Adding "maintain a kubernetes cluster" to its responsibilities is quite something, compared to "run docker-compose". Users are always free to run a kubernetes cluster of their own and add the NAS' storage via NFS (or some other StorageClass). | 15:26:13 |
aeonremnant | In reply to @sudoer:beeper.com
https://forums.truenas.com/t/the-future-of-electric-eel-and-apps/5409
K3s getting pulled out in favor of docker-compose Objective good honestly. TrueNAS was never the platform to have Kubernetes in it. Went against every design goal. | 15:27:03 |
DoofusCanadensis | okay, I'm confused about k3s and nodePort ranges. | 17:12:13 |
DoofusCanadensis | I'm trying to install homeassistant in my cluster and it wants to be on port 8123 (I assume this is a hard requirement for devices). But k3s says I'm only allowed 30000-32767 with no means to change that | 17:13:17 |
DoofusCanadensis | well, sort of. an older version of k3s might have let you change the range | 17:13:37 |
DoofusCanadensis | is nodePort ranges no longer modifiable? should I instead look at a dedicated machine for this homeassistant? | 17:14:16 |
Szwendacz | I use latest k3s and i use custom range | 17:14:45 |
Szwendacz | parameter --service-node-port-range 0-32767 | 17:15:11 |
DoofusCanadensis | okay, let me try this again | 17:15:32 |
DoofusCanadensis | ah, had to put it in the config. the one page I read had me add it to a node-args label | 17:30:07 |
26 Jul 2024 |
Rens Houben | Is there a way to dump the entire cluster configuration as a manifest so as to look it over and/or restore it from backup at a later date? | 06:39:13 |
effendy | Well, even if you could, you'd still also need to back up etcd to be able to restore the cluster to a certain state. | 08:05:09 |
| okdeer joined the room. | 08:42:15 |
Szwendacz | In reply to @rhouben:vac.systemec.nl Is there a way to dump the entire cluster configuration as a manifest so as to look it over and/or restore it from backup at a later date? I would say, that cluster deployment should be automated/orchestrated (?) | 08:57:39 |
Rens Houben | It should and currently mostly is, but I'm still finding my way and trying to figure out how to make things work, and now that I've gotten something to work I'd like a look at the entire config so I can trace my steps and see which parts actually worked. | 08:58:56 |