Implement pod resource metrics as described in KEP 1916. The new
`/metrics/resources` endpoint is exposed on the active scheduler
and reports kube_pod_resources* metrics that present the effective
requests and limits for all resources on the pods as calculated by
the scheduler and kubelet. This allows administrators using the
system to quickly perform resource consumption, reservation, and
pending utilization calculations when those metrics are read.
Because metrics calculation is on-demand, there is no additional
resource consumption incurred by the scheduler unless the endpoint
is scraped.
Allow a fast approximate conversion to float64 from quantity (which
is an arbitrary precision integral value). Will return +Inf/-Inf if
a float64 is not sufficient to represent the current value.
A suite of e2e tests was created for Topology Manager
so as to test pod scope alignment feature.
Co-authored-by: Pawel Rapacz <p.rapacz@partner.samsung.com>
Co-authored-by: Krzysztof Wiatrzyk <k.wiatrzyk@samsung.com>
Signed-off-by: Cezary Zukowski <c.zukowski@samsung.com>
Pod object is more flexible to use and construct
* Update TestGetTopologyHints() to work according to new test cases
* Update topologyHintTestCase{} to include proper field
Signed-off-by: Krzysztof Wiatrzyk <k.wiatrzyk@samsung.com>
* Extract common tests cases that will be used for both GetTopologyHints()
and GetPodTopologyHints()
* Extract machineInfo as it will be used for both functions as well
Signed-off-by: Krzysztof Wiatrzyk <k.wiatrzyk@samsung.com>
* Add topologyScopeName parameter to NewManager().
* Add scope interface and structure that implement common logic
* Add pod scope & container scopes
* Add pod lifecycle functions
Co-authored-by: sw.han <sw.han@samsung.com>
Signed-off-by: Krzysztof Wiatrzyk <k.wiatrzyk@samsung.com>
since we added tests to check connectivity against pods with
hostNetwork: true, there is the possibility that those pods
fail to run because the port is being used in the host.
Current test were using port 8080,8081 and 8082 that are commonly
used in hosts for other applications.
If the service is not ready after a certain time, and we are using
Pods with hostNetwork: true we assume that there is a conflict
and skip this test.
This PR changes the buckets of the
kubelet_runtime_operation_duration_seconds metric to be
metrics.ExponentialBuckets(.005, 2.5, 14) in order to
allow debugging image pull times. Right now the biggest bucket is 10
seconds, which is an ordinary time frame to pull an image, making the
metric useless for the aforementioned usecase.
Add a separate in-tree gcepd driver for windows cluster because it does
not support certain features as Linux driver.
Change-Id: I2fca86b3f32f17db7703c46a36944d9ee51f355f
