Build analytics on Hydra builds #86

Open
opened 2024-07-31 15:41:58 +00:00 by raito · 1 comment
Owner

Currently, we don't really know beyond what analytics Hydra provides on a certain attribute path in a jobset whether a certain derivation makes use of:

  • multiple cores
  • single core performance
  • fast IO
  • fast network

To better inform scheduling in the future, we can start learning this information. This is also useful to guide buying decisions for hardware upgrades.

Currently, we don't really know beyond what analytics Hydra provides on a certain attribute path in a jobset whether a certain derivation makes use of: - multiple cores - single core performance - fast IO - fast network To better inform scheduling in the future, we can start learning this information. This is also useful to guide buying decisions for hardware upgrades.
Owner

May I ask if you have something in mind around that already? Sounds like cgroup accounting on the builder side for jobs might be useful (still a pending CL IIRC?), right? And this is something Hydra could gather already I guess.

May I ask if you have something in mind around that already? Sounds like cgroup accounting on the builder side for jobs might be useful (still a pending CL IIRC?), right? And this is something Hydra could gather already I guess.
Sign in to join this conversation.
No milestone
No project
No assignees
2 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference: the-distro/infra#86
No description provided.