rollup_spec failing

[cben]

@yaacov @agrare kubernetes provider specs started failing 1–2 days ago:
https://travis-ci.org/ManageIQ/manageiq-providers-kubernetes/jobs/451822995
It's something related to metrics so I'm wondering if ManageIQ/manageiq#18159 merged 2 days ago might be related?

Failures:
  1) ManageIQ::Providers::Kubernetes::ContainerManager::Refresher checks pods running sequentially are not being multiplicated in project rollup
     Failure/Error:
       expect(project_rollup).to(
         have_attributes(
           :cpu_usage_rate_average     => 100
           :derived_memory_used        => 3072.0
           :derived_vm_numvcpus        => 12
           :derived_memory_available   => nil
           :mem_usage_absolute_average => nil
         )
       )
       expected #<MetricRollup id: 64000000000007, timestamp: "2012-09-01 00:00:00", capture_interval: nil, resource_...up_create_rate: 0, stat_container_group_delete_rate: 0, stat_container_image_registration_rate: nil> to have attributes {:cpu_usage_rate_average => 100, :derived_memory_used => 3072.0, :derived_vm_numvcpus => 12, :derived_memory_available => nil, :mem_usage_absolute_average => nil} but had attributes {:cpu_usage_rate_average => 100.0, :derived_memory_used => 3072.0, :derived_vm_numvcpus => 12.0, :derived_memory_available => 3072.0, :mem_usage_absolute_average => 100.0}
       Diff:
       @@ -1,6 +1,6 @@
       -:cpu_usage_rate_average => 100
       -:derived_memory_available => nil
       +:cpu_usage_rate_average => 100.0
       +:derived_memory_available => 3072.0
        :derived_memory_used => 3072.0
       -:derived_vm_numvcpus => 12
       -:mem_usage_absolute_average => nil
       +:derived_vm_numvcpus => 12.0
       +:mem_usage_absolute_average => 100.0
     Shared Example Group: "kubernetes rollup tests" called from ./spec/models/manageiq/providers/kubernetes/container_manager/metrics_capture/rollup_spec.rb:422
     # ./spec/models/manageiq/providers/kubernetes/container_manager/metrics_capture/rollup_spec.rb:253:in `block (2 levels) in <top (required)>'
  2) ManageIQ::Providers::Kubernetes::ContainerManager::Refresher check project rollup can handle partial hour
     Failure/Error:
       expect(project_rollup).to(
         have_attributes(
           :derived_memory_used        => 2048.0
           :derived_vm_numvcpus        => 12
           :derived_memory_available   => nil
           :mem_usage_absolute_average => nil
         )
       )
       expected #<MetricRollup id: 64000000000016, timestamp: "2012-09-01 00:00:00", capture_interval: nil, resource_...up_create_rate: 0, stat_container_group_delete_rate: 0, stat_container_image_registration_rate: nil> to have attributes {:derived_memory_used => 2048.0, :derived_vm_numvcpus => 12, :derived_memory_available => nil, :mem_usage_absolute_average => nil} but had attributes {:derived_memory_used => 2048.0, :derived_vm_numvcpus => 12.0, :derived_memory_available => 3072.0, :mem_usage_absolute_average => 66.6666666666667}
       Diff:
       @@ -1,5 +1,5 @@
       -:derived_memory_available => nil
       +:derived_memory_available => 3072.0
        :derived_memory_used => 2048.0
       -:derived_vm_numvcpus => 12
       -:mem_usage_absolute_average => nil
       +:derived_vm_numvcpus => 12.0
       +:mem_usage_absolute_average => 66.6666666666667
     Shared Example Group: "kubernetes rollup tests" called from ./spec/models/manageiq/providers/kubernetes/container_manager/metrics_capture/rollup_spec.rb:422
     # ./spec/models/manageiq/providers/kubernetes/container_manager/metrics_capture/rollup_spec.rb:210:in `block (2 levels) in <top (required)>'