github/descheduler
1
0
Fork 0
mirror of https://github.com/kubernetes-sigs/descheduler.git synced 2026-08-20 13:06:24 +00:00
Code Issues Packages Projects Releases Wiki Activity
1,724 Commits 24 Branches 89 Tags
node-utilization-util-snapshot
Commit Graph
614 Commits
Author SHA1 Message Date
Jan Chaloupka b56794708d descheduler: wire the metrics collector with the framework handle 2024-11-05 21:13:27 +01:00
Jan Chaloupka b7b352780e LowNodeUtilization: test metrics based utilization 2024-11-05 21:11:33 +01:00
Jan Chaloupka 646a383b37 Get pod usage from the usage client 2024-11-05 14:07:59 +01:00
Jan Chaloupka ad18f41b66 Update actualUsageClient 2024-11-04 18:11:27 +01:00
Jan Chaloupka 80f9c0ada6 Separate usage client into a new file 2024-10-21 22:38:25 +02:00
Jan Chaloupka 3174107718 usageSnapshot -> requestedUsageClient 2024-10-21 22:22:38 +02:00
Jan Chaloupka 1f55c4d680 node utilization: abstract pod utilization retriever 2024-10-15 12:18:37 +02:00
Jan Chaloupka dc9bea3ede nodeutiliation: create a usage snapshot 2024-10-15 12:18:30 +02:00
Jan Chaloupka 89bd188a35 hnu: move static code from Balance under plugin constructor 2024-10-11 16:49:23 +02:00
Jan Chaloupka e3c41d6ea6 lnu: move static code from Balance under plugin constructor 2024-10-11 16:37:53 +02:00
Jan Chaloupka e0ff750fa7 Move default LNU threshold setting under setDefaultForLNUThresholds 2024-10-11 16:31:37 +02:00
Simon ScharfandGitHub 22d9230a67 Make sure dry runs sees all the resources a normal run would do (#1526) 
* generic resource handling, so that dry run has all the expected resource types and objects

* simpler code and better names

* fix imports
 2024-10-04 12:20:28 +01:00
Kubernetes Prow RobotandGitHub 0f1890e5cd Merge pull request #1480 from ingvagabund/omitempty-for-plugin-args 
Plugin args: tag arguments with omitempty to reduce the marshalled json size
 2024-09-02 12:00:56 +01:00
Jan Chaloupka 29c0a90998 TestPodEvictorReset: replace duplicates strategy with node taints to simplify the testing 2024-08-14 11:00:20 +02:00
Kubernetes Prow RobotandGitHub 640b675e86 Merge pull request #1484 from ingvagabund/test-descheduling-limits 
[unit test]: test descheduling limits
 2024-08-14 01:53:04 -07:00
Kubernetes Prow RobotandGitHub c0c26e762b Merge pull request #1483 from ingvagabund/dedup-framework-init 
tests: de-duplicate framework handle initialization
 2024-08-14 01:20:26 -07:00
Jan Chaloupka 91e5e06b5f [unit test]: test descheduling limits 2024-08-14 10:15:58 +02:00
Jan Chaloupka cbade38d23 [tests] de-duplicate framework handle initialization 2024-08-12 17:05:30 +02:00
Jan Chaloupka 1e0b1a9840 Remove descheduler/v1alpha1 type 2024-08-09 09:49:59 +02:00
Jan Chaloupka cb0c1b660d Plugin args: tag arguments with omitempty to reduce the marshalled json size 2024-08-06 15:20:18 +02:00
Amir Alavi a3146a1705 fix: minor version parsing in version compatibility check 
Signed-off-by: Amir Alavi <amiralavi7@gmail.com>
 2024-07-28 11:44:12 -04:00
Victor GonzalezandGitHub 55a0812ae6 skip eviction when pod creation time is below minPodAge threshold setting (#1475) 
* skip eviction when pod creation time is below minPodAge threshold setting

In the default initialization phase of the descheduler, add a new
constraint to not evict pods that creation time is below minPodAge
threshold.

Added value:

- Avoid crazy pod movement when the autoscaler scales up and down.

- Avoid evicting pods when they are warming up.

- Decreases the overall cost of eviction as no pod will be evicted
  before doing significant amount of work.

- Guard against scheduling. Descheduling loops in situations where
  the descheduler has a different node fit logic from scheduler,
  like not considering topology spread constraints.

* Use *time.Duration instead of uint for MinPodAge type

* Remove '(in minutes)' from default evictor configuration table

* make fmt

* Add explicit name for Duration field

* Use Duration.String()
 2024-07-26 05:59:21 -07:00
Adam Malcontenti-Wilson f23967a88e feat: add init and ephemeral container checks to PodLifeTime 2024-07-17 14:36:35 +10:00
Emin Aktas f8e128d862 refactor: replace k8s.io/utils/pointer with k8s.io/utils/ptr 
Signed-off-by: Emin Aktas <eminaktas34@gmail.com>
 2024-07-11 11:36:34 +03:00
Kubernetes Prow RobotandGitHub b614c8bc7c Merge pull request #1458 from ingvagabund/pod-evictor-thread-safe 
pod evictor: make it thread safe
 2024-07-10 04:48:26 -07:00
Kubernetes Prow RobotandGitHub 9b41edd382 Merge pull request #1460 from ingvagabund/context 
descheduler_test.go: initDescheduler: pass a new ctx with cancel inst…
 2024-07-09 11:24:11 -07:00
Jan Chaloupka bc60a058ef pod evictor: make it thread safe 
Currently, all the plugins are run in a sequence.
No plugin executes evictions in parallel within.
Yet, there's no guarantee a future plugin (e.g. a custom one)
will not attemp to evict pods in parallel.
 2024-07-09 12:39:08 +02:00
Kubernetes Prow RobotandGitHub 546a39e88c Merge pull request #1451 from zhifei92/support-total-pods-limit 
The descheduler supports limiting the total number of pods evicted per rescheduling cycle
 2024-07-09 03:35:17 -07:00
zhifei92 578086ca8e refactor: optimize error descriptions 2024-07-09 17:14:34 +08:00
Jan Chaloupka ea2eeccff4 descheduler_test.go: initDescheduler: pass a new ctx with cancel instead of returning a cancel 2024-07-09 09:19:59 +02:00
zhifei92 e60f525ec6 feat: support MaxNoOfPodsToEvictTotal 2024-07-09 14:00:27 +08:00
Jan Chaloupka 3362fec7b0 Define initDescheduler for further use 
Can be used by other tests executing individual descheduling cycle
explicitly.
 2024-07-08 17:28:51 +02:00
Jan Chaloupka f240648df2 Set OwnerReferences through GetReplicaSetOwnerRefList 2024-07-08 17:10:07 +02:00
Jan Chaloupka a818c01832 Use v1alpha2 descheduling policy 2024-07-08 17:04:53 +02:00
Jan Chaloupka 44b59f9b1d initPluginRegistry as a single way to register all plugins in testing 2024-07-08 16:55:27 +02:00
Kubernetes Prow RobotandGitHub 9d16c28f43 Merge pull request #1456 from ingvagabund/limit-exceeded-to-error 
PodEvictor: turn an exceeded limit into an error
 2024-07-08 06:25:06 -07:00
googs1025 db0df6c6ca fix: add info for error return 2024-07-07 21:16:52 +08:00
Jan Chaloupka 18d0e4a540 PodEvictor: turn an exceeded limit into an error 
When checking for node limit getting exceeded the pod eviction
never fails. Thus, ignoring the metric reporting when a pod fails
to be evicted due to node limit constrains.

The error also allows plugin to react on other limits getting
exceeded. E.g. the limit on the number of pods evicted per namespace.
 2024-07-06 20:14:43 +02:00
Kubernetes Prow RobotandGitHub 7657345079 Merge pull request #1452 from a7i/defaultevictor-reinit 
fix: indexer cache error when default evictor is re-initialized
 2024-07-06 04:30:53 -07:00
Amir Alavi e26f6429a2 feat: pod evictor options 2024-07-05 21:40:55 -04:00
Amir Alavi 7ab36daaec fix: indexer cache error when default evictor is re-initialized 
Signed-off-by: Amir Alavi <amiralavi7@gmail.com>
 2024-06-28 00:34:55 -04:00
Kubernetes Prow RobotandGitHub f2be3fd414 Merge pull request #1436 from fanhaouu/feat-return-node-fit-error 
return node fit error in advance
 2024-06-27 10:54:41 -07:00
googs1025 cfa6845a19 refactor: PodMatchNodeSelector method 2024-06-27 22:10:41 +08:00
Hao Fan 8a2b2eb37c return node fit error in advance 2024-06-26 17:18:32 +08:00
Kubernetes Prow RobotandGitHub 972d28108a Merge pull request #1447 from ingvagabund/refactorings 
PodEvictor: refactoring and preparation for eviction requests
 2024-06-25 01:08:44 -07:00
Kubernetes Prow RobotandGitHub f294d953a3 Merge pull request #1445 from zhifei92/fix-unit-test 
Unit-test: add necessary ownerRef to the pod.
 2024-06-25 00:15:40 -07:00
Jan Chaloupka fadef326ff TestPodEvictorReset: check the dry mode evicts duplicated pods 2024-06-23 20:30:36 +02:00
Jan Chaloupka f5060adcd1 Move fake client from the cachedClient function 
Remove the fakeClient from cachedClient function so a different
fakeClient can be injected for testing purposes
 2024-06-23 19:59:46 +02:00
Jan Chaloupka 75880226c0 Set up the pod evictor only once 
Currently, the pod evictor is created during each descheduling cycle
to reset the internal counters and the fake client (in case a dry run is
configured). Instead, create the pod evictor once and reset only what's
needed. So later on the pod evictor can be extended with e.g. a cache
keeping the track of eviction requests that are still in progress and
required more than a single descheduling cycle to complete.
 2024-06-23 19:24:27 +02:00
Jan Chaloupka 0901cb18bf NewPodEvictor: drop nodes parameter 2024-06-22 15:08:00 +02:00