Open aarontp opened 7 years ago
From a GKE autoscaling perspective, we can consider:
Also handy resource: https://medium.com/google-cloud/scale-your-kubernetes-cluster-to-almost-zero-with-gke-autoscaler-9c78051cbf40
Tagging 'celery' because I think we can use built-in functionality for this
We can add different worker pools that are created to match individual job types (e.g. so Plaso can have worker nodes that have more cpu than workers for other job types). Right now it's one global worker pool.