Jan Chaloupka 392fc6668f Jittering periods of some kubelet's sync loops: ...

- pod_workers: pod syncing
- prober workers: container syncing

In order to synchronize the current state of Kubernetes's objects (e.g. pods, containers, etc.),
periodic synch loops are run. When there is a lot of objects to synchronize with,
loops increase communication traffic. At some point when all the traffic interfere cpu usage curve
hits the roof causing 100% cpu utilization.

To distribute the traffic in time, some sync loops can jitter their period in each loop
and help to flatten the curve.

2016-02-09 15:22:22 +01:00

..

cadvisor

Kubelet Metrics Summary Api Implementation

2016-02-04 14:05:28 -08:00

client

Merge pull request #19334 from resouer/network

2016-01-21 10:26:13 -08:00

cm

Rewritte util.* -> wait.* wherever reasonable

2016-02-07 12:02:20 +01:00

config

Rewritte util.* -> wait.* wherever reasonable

2016-02-07 12:02:20 +01:00

container

support optional envvars for pod-infra-container

2016-02-08 13:48:25 +00:00

custommetrics

Set Docker custom metric label for cAdvisor if custom metric volume is mounted

2016-02-05 12:00:02 +01:00

dockertools

support optional envvars for pod-infra-container

2016-02-08 13:48:25 +00:00

envvars

…

leaky

…

lifecycle

Remove GetAPIPodStatus usage

2016-02-02 23:41:55 +00:00

metrics

Supply volume fs metrics to server/stats/handler.go

2016-02-05 16:00:24 -08:00

network

rename release_1_2 to internalclientset

2016-02-05 14:02:28 -08:00

pleg

Rewritte util.* -> wait.* wherever reasonable

2016-02-07 12:02:20 +01:00

pod

rename release_1_2 to internalclientset

2016-02-05 14:02:28 -08:00

prober

Jittering periods of some kubelet's sync loops:

2016-02-09 15:22:22 +01:00

qos

hardcode supportedComputeResources

2015-11-19 13:20:02 +08:00

rkt

Remove ConvertPodStatusToAPIPodStatus from runtime interface

2016-02-05 19:19:46 +00:00

server

Rewritte util.* -> wait.* wherever reasonable

2016-02-07 12:02:20 +01:00

status

Rewritte util.* -> wait.* wherever reasonable

2016-02-07 12:02:20 +01:00

types

Implement node Allocatable

2016-01-12 17:00:43 -08:00

util

extend fake clock

2016-02-01 15:36:15 -08:00

container_bridge.go

Allow non-masquerade-cidr to be passed to the kubelet

2016-01-21 15:18:31 -05:00

disk_manager_test.go

Treat first-seen, non-pending pods as updates

2016-01-08 17:21:43 -08:00

disk_manager.go

Treat first-seen, non-pending pods as updates

2016-01-08 17:21:43 -08:00

doc.go

…

fake_pod_workers.go

Remove RuntimeCache from sync path

2016-02-01 21:32:41 -08:00

flannel_helper.go

Turn flannel off by default

2015-11-25 09:43:48 +01:00

image_manager_test.go

extend fake clock

2016-02-01 15:36:15 -08:00

image_manager.go

Rewritte util.* -> wait.* wherever reasonable

2016-02-07 12:02:20 +01:00

kubelet_test.go

Rewritte util.* -> wait.* wherever reasonable

2016-02-07 12:02:20 +01:00

kubelet.go

support optional envvars for pod-infra-container

2016-02-08 13:48:25 +00:00

networks.go

rename release_1_2 to internalclientset

2016-02-05 14:02:28 -08:00

oom_watcher_test.go

…

oom_watcher.go

Replace runtime reference by pkg

2016-02-01 21:06:44 +08:00

pod_workers_test.go

Remove RuntimeCache from sync path

2016-02-01 21:32:41 -08:00

pod_workers.go

Jittering periods of some kubelet's sync loops:

2016-02-09 15:22:22 +01:00

reason_cache_test.go

Add unit test for reason cache

2016-01-30 00:25:50 -08:00

reason_cache.go

Add reason cache

2016-01-30 00:25:50 -08:00

root_context_linux.go

Bump cadvisor to fix interface stats bugs & improve performance

2015-12-21 17:07:21 +00:00

root_context_unsupported.go

…

runonce_test.go

Remove GetAPIPodStatus usage

2016-02-02 23:41:55 +00:00

runonce.go

Remove GetAPIPodStatus usage

2016-02-02 23:41:55 +00:00

runtime.go

Send PodCIDR to network plugins as an event

2016-02-03 10:08:07 -06:00

util.go

…

volume_manager.go

…

volumes.go

Merge pull request #19741 from pwittrock/syncfsmetrics

2016-02-05 17:54:22 -08:00