#spacelift (2023-04)
2023-04-04
Introducing Spacelift Self-Hosted We are excited to announce that now it is possible to self-host an entire, fully-functional Spacelift platform within your AWS cloud.
2023-04-26
What Does Terraform Refresh Command Do – Overview Learn what the Terraform refresh is and why it is recommended not to use this command directly.
@Erik Osterman (Cloud Posse) has CloudPosse built any tooling to manage the number of concurrent private workers running?
Kind of? We use a primitive scaling technique increasing the number of runners in an auto scale group when there’s active utilization, and shrinking it when utilization drops. It’s conservative in both directions. Utilization is defined like 5-10% CPU; idle Is 0-1%. Then we set a ceiling and floor.
Ideally workers would be on demand.
This has worked well, but customers can get dinged pretty badly if they use a lot of runners, as the cost per runner adds up. We manage security hub and guard duty in terraform in 17+ regions in 20+ accounts. When nightly drift detection kicks off, it spawns thousands of runs. If workers didn’t scale up, spacelift would be backlogged for hours and hours.
yeah. We have lots of worker pools and want to manage the number of active workers intelligently across all of them
That makes sense, and the kubernetes implementation leaves a lot to be desired.
It should be implemented as a controller and be able to scale workers to zero
Cc @marcinw
I’m not going to go into implementation details on a public channel but if you’re interested in the preliminary design I will loop you in as a design partner for the feature?
Great to hear! I know some customers would be interested in that. Will reach out.