Pulled from #93 to collect data before deciding on final cost formula.
Adds the following columns:
nodes: AWS zone (zone)
nodes: Instance capacity type (capacity_type)
jobs: Cost of the instance during the lifetime of the job (job_cost_instance)
The job_cost_instance calculation is made by averaging the value of karpenter_cloudprovider_instance_type_offering_price_estimate during the lifetime of the node and multiplying by the duration of the build job.
This is not a cost per job metric. Use information like cpu_mean, mem_mean, etc to calculate the cost of the job in combination with job_cost_instance.
tested with dev/bulk_collect.py and verified that large migrations work correctly on the prod and staging db
Pulled from #93 to collect data before deciding on final cost formula.
Adds the following columns:
zone)capacity_type)job_cost_instance)The
job_cost_instancecalculation is made by averaging the value ofkarpenter_cloudprovider_instance_type_offering_price_estimateduring the lifetime of the node and multiplying by the duration of the build job.This is not a cost per job metric. Use information like cpu_mean, mem_mean, etc to calculate the cost of the job in combination with
job_cost_instance.tested with
dev/bulk_collect.pyand verified that large migrations work correctly on the prod and staging db