| 3 May 2025 |
 tea | hi all, I noticed that the collection of fastly metrics at https://github.com/NixOS/infra/tree/main/metrics/fastly seems to have stopped since february. not sure who maintains these, but I think this is valuable data that would be nice to have. | 16:10:38 |
| tea | * hi all, I noticed that the collection of fastly metrics at https://github.com/NixOS/infra/tree/main/metrics/fastly seems to have stopped in february. not sure who maintains these, but I think this is valuable data that would be nice to have. | 16:12:05 |
 hexa | yeah, stuck here https://github.com/NixOS/infra/pull/388 | 16:13:29 |
| tea | ah, okay | 16:14:50 |
|  @jh-devv:matrix.org joined the room. | 18:25:21 |
| 4 May 2025 |
|  @federicodschonborn:matrix.org joined the room. | 01:31:47 |
| @jh-devv:matrix.org | What are some other milestones for infra other than some changes in regards for security? | 16:47:05 |
 K900 | There are no Milestones | 16:49:50 |
| K900 | There's no formal planning | 16:49:54 |
| K900 | The big issues would be cache GC and scaling Hydra | 16:50:00 |
 Mic92 | It's unclear to me where this was actually deployed | 16:51:07 |
 Jeremy Fleischman (jfly) | According to https://github.com/NixOS/infra/pull/388, it sounds like it was running in Eelco's closet | 16:52:08 |
| hexa | Redacted or Malformed Event | 16:52:11 |
| hexa | https://github.com/NixOS/infra/tree/main/metrics/fastly#reports | 16:52:33 |
| @jh-devv:matrix.org | In reply to @k900:0upti.me
There's no formal planning I guess there are weekly meetups but well, more like current issues
Mainly seems like the fastly metrics are the only big one rn | 17:13:13 |
|  sinan changed their profile picture. | 18:36:26 |
| sinan changed their profile picture. | 18:38:37 |
| sinan | 18:38:45 |
| sinan | 18:38:48 |
| sinan | 18:39:38 |
| sinan | 18:41:08 |
|  ch joined the room. | 18:53:34 |
|  Nikke joined the room. | 21:27:55 |
| 5 May 2025 |
|  @shon2929:matrix.org joined the room. | 01:44:54 |
| @shon2929:matrix.org left the room. | 01:45:33 |
 leona | it feels that something is broken with hydra. eval for trunk-combined takes >1h20 (still running), build queue is super small but the scheduler and ingestion is empty according to grafana | 09:32:32 |
| K900 | That's normal | 09:33:19 |
| K900 | It's doing a single transaction insert of all the builds into the database first | 09:33:34 |
| K900 | And only then the queue runner starts ingesting | 09:33:44 |
| K900 | Hydra is very sane software | 09:33:59 |
