| 23 Aug 2022 |
 bryton | I'm currently trying to dust off kubenix to migrate my old helmfile/Pulumi mess into nix. Current focus is documentation and a CLI; note that I'm still learning nix so I'm bound to make some blunders here and there if anyone wants to stop by and correct me (or just make suggestions in general) 😁 | 03:57:32 |
|  implr left the room. | 10:08:01 |
|  luxus joined the room. | 19:27:36 |
| 24 Aug 2022 |
|  underpantsgnome changed their display name from tinybronca to underpantsgnome. | 23:07:14 |
| 30 Aug 2022 |
|  aru joined the room. | 14:42:53 |
| aru left the room. | 14:52:50 |
| 8 Sep 2022 |
|  aharris joined the room. | 22:13:42 |
| 15 Sep 2022 |
|  m_algery joined the room. | 12:18:55 |
| m_algery left the room. | 12:18:58 |
| 20 Sep 2022 |
|  raphi joined the room. | 12:10:50 |
 hexa | how much am I going to hate myself for setting up an etcd cluster with only two nodes? | 13:20:58 |
 Arian | There is no point | 13:22:00 |
| Arian | A single etcd node will give you the same guarantees | 13:22:22 |
| Arian | As only the leader will process writes and reads. And there is no failover detection when the leader dies | 13:22:43 |
| Arian | As there is no consensus | 13:22:50 |
| Arian | Either have 3 or 1 | 13:22:58 |
| hexa | alright, thanks | 13:23:05 |
| Arian | (or 5) | 13:23:14 |
| hexa | yeah, so uneven | 13:23:50 |
| Arian | Etcd writes don't scale beyond the performance of a single node by the way | 13:23:51 |
| hexa | * yeah, so odd numbers | 13:23:54 |
| Arian | Reads you can scale | 13:23:57 |
| hexa | this is about using dex with an ldap syncrepl pair | 13:24:10 |
| Arian | Just run a single node | 13:24:21 |
| hexa | I don't think we're going to need much throughput | 13:24:22 |
| hexa | ok | 13:24:36 |
| hexa | thanks | 13:24:38 |
| 21 Sep 2022 |
|  superherointj joined the room. | 20:45:36 |
|  Felix Andreas joined the room. | 22:16:14 |
| 22 Sep 2022 |
|  vivian joined the room. | 22:24:01 |
