content large_stringlengths 3 20.5k | url large_stringlengths 53 192 ⌀ | branch large_stringclasses 4
values | source large_stringclasses 50
values | embeddings listlengths 384 384 | score float64 -0.21 0.65 |
|---|---|---|---|---|---|
\*\*Node Selection\*\*: Will prioritize empty nodes for termination. Incase of non-empty nodes we will pick at random. In future iterations, we will add intelligent algorithms for selecting which nodes to scale down - \*\*Termination Behavior\*\*: Will gracefully terminate nodes to meet desired size respecting PodDisru... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/static-capacity.md | main | karpenter | [
-0.0513434037566185,
0.04051431640982628,
0.036425553262233734,
0.07754015177488327,
0.053698450326919556,
-0.07019124180078506,
-0.05344614014029503,
-0.04198905825614929,
-0.04682169109582901,
0.04753481224179268,
-0.07075612992048264,
-0.0018727175192907453,
0.0169078279286623,
-0.01027... | 0.158673 |
# Karpenter NoSchedule Taint [v0.32.0+] ## Overview Karpenter uses the `node.kubernetes.io/unschedulable:NoSchedule` taint to prevent pods from scheduling to nodes being deprovisioned. The `node.kubernetes.io/unschedulable` taint is well-known in Kubernetes and may be leveraged or relied on by other controllers or appl... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/no-schedule-taint.md | main | karpenter | [
-0.022335687652230263,
-0.0075574032962322235,
0.0513576865196228,
0.004987237509340048,
0.05192947015166283,
-0.015280750580132008,
-0.016640909016132355,
-0.009088587015867233,
0.07514587044715881,
0.0878906324505806,
0.05241779237985611,
-0.016589509323239326,
-0.03551710397005081,
-0.0... | 0.280866 |
this conversation and, instead, taking an opinionated stance in Karpenter that we can choose to evolve later. ### How severe is this issue currently? At the moment, no user has reported this issue, likely since its effect and recovery are mitigated by how Karpenter deprovisions serially, executing one action at a time.... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/no-schedule-taint.md | main | karpenter | [
-0.06219429150223732,
-0.07901133596897125,
0.08546113967895508,
0.05856885388493538,
0.012755549512803555,
0.005571047309786081,
-0.05599850416183472,
0.005142523441463709,
0.04321447014808655,
0.08858145028352737,
0.016257137060165405,
0.013969996012747288,
-0.016104014590382576,
0.00406... | 0.100044 |
# Node Auto Repair ## Problem Statement Nodes can experience failure modes that cause degradation to the underlying hardware, file systems, or container environments. Some of these failure modes are surfaced through the Node object such as network unavailability, disk pressure, or memory pressure, while others are not ... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/node-repair.md | main | karpenter | [
-0.032419245690107346,
0.042551908642053604,
0.08668036758899689,
0.07274018973112106,
0.042415350675582886,
-0.04795704036951065,
-0.005477115977555513,
-0.0009643298108130693,
-0.01197383925318718,
0.08514105528593063,
-0.029867593199014664,
-0.025428328663110733,
0.01862317882478237,
-0... | 0.209335 |
connection reason: Network is not available status: "False" type: NetworkUnavailable ... - The Node here will be eligible for node repair after at `2024-11-01T15:12:48Z` --- apiVersion: v1 kind: Node metadata: ... status: condition: - lastHeartbeatTime: "2024-11-01T16:29:49Z" lastTransitionTime: "2024-11-01T15:02:48Z" ... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/node-repair.md | main | karpenter | [
-0.05507335439324379,
0.010935352183878422,
0.04794963076710701,
-0.01769426092505455,
-0.02477513812482357,
-0.0176081620156765,
-0.08427022397518158,
-0.028964534401893616,
0.0056511396542191505,
0.08329544961452484,
-0.0021203362848609686,
-0.046727124601602554,
-0.05253920704126358,
-0... | 0.121474 |
# Forceful Expiration - v0.37+ Users need to enforce a maximum node lifetime to be compliant with security requirements. Cluster administrators that orchestrate Karpenter installations on clusters need a way to enforce this natively within Karpenter. In v0.37 and below, Karpenter respects NDBS, `do-not-disrupt` annotat... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/forceful-expiration.md | main | karpenter | [
-0.031958431005477905,
0.014960544183850288,
0.0487283319234848,
-0.001493919873610139,
-0.012135823257267475,
-0.024531178176403046,
-0.03234446793794632,
-0.029074233025312424,
0.02913765422999859,
0.06458576023578644,
0.013891304843127728,
-0.0015321745304390788,
-0.014779327437281609,
... | 0.209295 |
The only difference in the diagram below is Expiration is moved to \*\*T1 (S1 → S2)\*\*.  ## 🔑 Adding Forceful Expiration ### Option 1 - Make Expiration Forceful - Recommended Revert Expiration’s semantic to be \*\*Forceful\*\*, bypassing past NDBs and `do-not-disrupt`, starting gr... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/forceful-expiration.md | main | karpenter | [
-0.0025731620844453573,
0.013403939083218575,
0.06380211561918259,
0.019880564883351326,
0.06971623003482819,
-0.012105385772883892,
-0.032821692526340485,
0.05328261852264404,
0.014524216763675213,
0.021835336461663246,
0.02777881547808647,
0.06638386845588684,
0.04896526783704758,
-0.048... | 0.1789 |
# Deprovisioning - Drift This will be a living document, where Karpenter details the key design choices for Drift, Karpenter's mechanism to assert created machines match the corresponding CRDs. At its inception, this doc only included AWS fields. Future cloud providers should add to this document to add to the table be... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/drift.md | main | karpenter | [
-0.08353722095489502,
0.04989182576537132,
0.08519965410232544,
-0.015213582664728165,
0.047646407037973404,
-0.0032083094120025635,
0.002214432694017887,
-0.02041620947420597,
0.04702772945165634,
0.06720399856567383,
-0.006743457168340683,
-0.12385917454957962,
0.06373842805624008,
-0.04... | 0.125 |
Fields - | --- | --- | --- | | Startup Taints | x | | | | Taints | x | | | | Labels | x | | | | Annotations | x | | | | Node Requirements | | x | | | Kubelet Configuration | x | | | | Weight | | | x | | Limits | | | x | | Consolidation | | | x | | TTLSecondsUntilExpired | | | x | | TTLSecondsAfterEmpty | | | x | | - AW... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/drift.md | main | karpenter | [
-0.08234192430973053,
0.017462654039263725,
0.005440582055598497,
-0.00899337325245142,
0.01340454537421465,
-0.01896749809384346,
0.026404812932014465,
0.0051429481245577335,
0.034962136298418045,
0.07092935591936111,
0.040692999958992004,
-0.10268285125494003,
0.0779106467962265,
-0.0573... | 0.15663 |
systems to modify nodes and instances, but keep the drift automation of rolling out changes to settings to their fleet. Furthermore, triggering Drift for Static Fields only on CRD changes allows users to safely control when their fleet will be refreshed. As a reference, this is how Kubernetes Deployments work. A user c... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/drift.md | main | karpenter | [
-0.02933535911142826,
-0.004286628682166338,
0.07049661874771118,
0.01247828546911478,
-0.0030769065488129854,
0.002958450233563781,
0.01901526190340519,
-0.023189600557088852,
0.08425840735435486,
0.04589398577809334,
0.014841292053461075,
-0.030463140457868576,
0.029071195051074028,
-0.0... | 0.154959 |
# NodeClaim Termination Grace Period ## Motivation Users are requesting the ability to control how long karpenter will wait for the deletion of an individual node to complete before forcefully terminating the node regardless of pod status. [#743](https://github.com/kubernetes-sigs/karpenter/issues/743). This supports t... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/nodeclaim-termination-grace-period.md | main | karpenter | [
-0.03801915422081947,
0.05573858320713043,
0.08027858287096024,
0.00906766764819622,
0.04277044162154198,
0.0046822489239275455,
-0.0004226916062179953,
-0.017277099192142487,
0.08790731430053711,
0.09318095445632935,
0.0029038444627076387,
0.021587567403912544,
-0.006789310369640589,
-0.0... | 0.229641 |
would not be able to take advantage of this feature. ## Validation/Defaults The `terminationGracePeriod` fields accepts a common duration string which defaults to a value of `nil`. Omitting the field results in the default nil value instructing the controller to wait indefinitely for pods to drain gracefully, maintaini... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/nodeclaim-termination-grace-period.md | main | karpenter | [
0.024079585447907448,
0.06784680485725403,
0.06504552811384201,
0.055240631103515625,
0.01902173087000847,
-0.017671819776296616,
-0.04538044333457947,
-0.023208368569612503,
0.06642867624759674,
0.03506897762417793,
-0.00825438555330038,
-0.056930504739284515,
-0.05003119260072708,
-0.039... | 0.21298 |
Each of these pods requires the full 10m to terminate gracefully. 3. The Deployment has an associated PDB with maxUnavailable 1 requiring that only one of these pods be disrupted at a time. 1. The cluster admin has configured NodePool.spec.disruption.terminationGracePeriod at 15m. 2. A full node reconcilliation has sta... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/nodeclaim-termination-grace-period.md | main | karpenter | [
0.033487703651189804,
0.01964537613093853,
0.08681860566139221,
0.0077735660597682,
0.05545641481876373,
-0.017737977206707,
-0.05047914385795593,
-0.01702682301402092,
0.02448800951242447,
0.06949449330568314,
0.01122735720127821,
-0.0001942659291671589,
-0.024138424545526505,
-0.04486088... | 0.130084 |
delete, even if eviction is supported. This will bypass checking PodDisruptionBudgets, use with caution. ``` | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/nodeclaim-termination-grace-period.md | main | karpenter | [
0.041798561811447144,
0.03184463083744049,
0.029529206454753876,
-0.05993269383907318,
0.04505248740315437,
-0.038790423423051834,
0.005385002586990595,
-0.12257479131221771,
0.09116005152463913,
0.044123873114585876,
0.08081436902284622,
0.0662064179778099,
0.018588777631521225,
0.0172940... | 0.001308 |
# Karpenter v1 Roadmap ## Overview Karpenter released the beta version of its APIs and features in October 2023. The intention behind this beta was that we would be able to determine the final set of changes and feature adds that we wanted to add to Karpenter before we considered Karpenter feature-complete. The list be... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/v1-roadmap.md | main | karpenter | [
-0.060823872685432434,
0.004671947564929724,
0.12058387696743011,
-0.006607883144170046,
0.02252955548465252,
-0.012631629593670368,
-0.08119229227304459,
-0.0006941711180843413,
-0.08764459937810898,
0.041797470301389694,
-0.029890194535255432,
-0.05068736895918846,
-0.08118641376495361,
... | 0.158463 |
on the node on startup, this process or registration can break scheduling compatibility. For instance, a pod may not tolerate a specific taint that was configured in the NodeClaim template, but the taint did not appear on the node immediately on startup, meaning that the pod was able to successfully schedule when it sh... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/v1-roadmap.md | main | karpenter | [
-0.06557396054267883,
0.016090888530015945,
0.096908338367939,
0.034524526447057724,
0.08952760696411133,
0.02329646609723568,
-0.027658823877573013,
-0.045086488127708435,
0.06273883581161499,
0.05975660681724548,
-0.020359128713607788,
-0.03224441781640053,
-0.0070388903841376305,
0.0165... | 0.130683 |
Introduce ConsolidateAfter for Consolidation Controls \*\*Issue Ref(s):\*\* https://github.com/kubernetes-sigs/karpenter/issues/735 \*\*Category\*\*: Breaking, Stability Karpenter currently evaluates all forms of disruption in a synchronous loop, starting with expiration, drift, emptiness, and considering consolidation... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/v1-roadmap.md | main | karpenter | [
-0.04042438790202141,
-0.03405557572841644,
0.048364508897066116,
0.06447799503803253,
0.00008797546615824103,
-0.04256989434361458,
-0.01053929515182972,
-0.05749557167291641,
0.06516806781291962,
0.05176933854818344,
0.03734878823161125,
0.012400339357554913,
0.013062478974461555,
-0.069... | 0.153302 |
Ref:\*\* https://github.com/kubernetes-sigs/karpenter/issues/1051 \*\*Category:\*\* Breaking, Stability Karpenter needs to stabalize a set of metrics, status conditions, and events that can be relied-upon for monitoring. The design for these metrics, status conditions and events will be added in a separate RFC. ### Nod... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/v1-roadmap.md | main | karpenter | [
-0.07944915443658829,
0.020238088443875313,
0.05288701504468918,
0.03746930882334709,
0.040042176842689514,
0.010640029795467854,
-0.0004940381622873247,
0.008425666019320488,
0.060001667588949203,
0.07725510001182556,
-0.03706180304288864,
-0.08327266573905945,
0.002195316832512617,
-0.04... | 0.216199 |
if users request it) ### Promoting Drift Feature to Stable \*\*Category:\*\* Stability Karpenter supported drift in its alpha state from v0.21-v0.32. During alpha, we worked to build out features and promoted drift to beta after releasing full drift support for all Karpenter configuration. Since releasing drift in beta... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/v1-roadmap.md | main | karpenter | [
-0.07121513783931732,
-0.04199625179171562,
0.09154807031154633,
0.04313797503709793,
0.02181571163237095,
-0.0073479413986206055,
0.010128351859748363,
-0.060145728290081024,
-0.06829886138439178,
0.02917299047112465,
-0.035016756504774094,
-0.067181296646595,
-0.05519130453467369,
0.0219... | 0.149327 |
# GTE and LTE Operators for Requirements ## Overview Karpenter currently supports `Gt` (greater than) and `Lt` (less than) operators for numeric requirements. These are strict inequalities, which creates an awkward user experience. For example, to require a minimum of 2 CPUs, users must specify `Gt 1` instead of the mo... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/gte-lte-operators.md | main | karpenter | [
-0.0156831294298172,
-0.00967162474989891,
0.04982290416955948,
-0.010995629243552685,
-0.05506521835923195,
-0.03172225505113602,
-0.02355613373219967,
0.07540600001811981,
-0.04247558116912842,
-0.015557346865534782,
-0.048762883991003036,
-0.07045171409845352,
0.0497785285115242,
0.0064... | 0.111733 |
# Disruption Controls - (in v1beta1+) ## Motivation Users have been asking for more control around the speed of Consolidation and the ability to restrict disruptions to their nodes. The demand for this is reflected by the +1s in [#1738](https://github.com/aws/karpenter/issues/1738). This is increasingly important as cl... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/disruption-controls.md | main | karpenter | [
0.005495655816048384,
-0.004009697120636702,
0.03101528249680996,
0.07616930454969406,
0.10992417484521866,
-0.020800765603780746,
-0.009742897003889084,
0.00180807092692703,
-0.013449609279632568,
0.04492304101586342,
0.00767249520868063,
-0.03773094713687897,
0.00844633486121893,
-0.0553... | 0.154983 |
disable scale down for a NodePool by setting this to `0`. Users must either omit both `Schedule` and `Duration` or set both of them, since `Schedule` and `Duration` are inherently linked. Omitting these two fields will be equivalent to an always active `Budget`. Users cannot define a seconds nodes in `Duration`, since ... | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/disruption-controls.md | main | karpenter | [
-0.011087684892117977,
0.04146308824419975,
-0.04239613190293312,
0.06433368474245071,
-0.012838338501751423,
-0.012365701608359814,
-0.03294742479920387,
-0.013616465032100677,
0.009538570418953896,
0.028517769649624825,
0.02932862751185894,
-0.07115304470062256,
-0.009013413451611996,
-0... | 0.112275 |
on the same spec. \* 👍 Karpenter already uses this mechanism which is done with `Provisioner → ProviderRef` / `NodePool -> NodeClass`. \* 👎 If we wanted to align all Disruption fields together, migrating the other `consolidateAfter` and `expireAfter` fields over to a new could confuse users. | https://github.com/kubernetes-sigs/karpenter/blob/main/designs/disruption-controls.md | main | karpenter | [
-0.08058565855026245,
-0.034010037779808044,
-0.002902927342802286,
0.01841481775045395,
0.0014598496491089463,
-0.014926455914974213,
-0.01365293562412262,
-0.027880415320396423,
0.022926107048988342,
0.08185974508523941,
0.09350955486297607,
-0.042054351419210434,
0.0012125499779358506,
... | 0.116547 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.