v0.2.0 release

What's Changed

• chore: release chart by @jwcesign in #142
• fix: fix telemetry flag parsing by @jwcesign in #143
• ci: add release image build by @jwcesign in #144
• add taints for node by @jxs1211 in #145
• chore: add tag to apg to enhance the SLA by @jwcesign in #151
• fix: give higher memory overhead percent by @jwcesign in #153
• fix: calculate the overhead with the maximum value by @jwcesign in #155
• fix: calculate the overhead correctly by @jwcesign in #158
• chore: use SecurityGroupIds instad of SecurityGroupId for ProvisionGroup by @Vacant2333 in #157
• release: release v0.1.2 by @jwcesign in #166
• typo: fix typo by @jwcesign in #168
• fix: get image status even it's expired by @jwcesign in #169
• doc: Update Wechat Group by @Aleegra in #171
• fix: fallback to OnDemand when Spot is not available by @Vacant2333 in #172
• chore: release patch version v0.1.3 by @jwcesign in #174
• typo: fix typo by @jwcesign in #175
• improve authorization by @daimaxiaxie in #152
• add nodeSelector for karpenter node selection by @jxs1211 in #176
• chore: handle vswitch not found case in getProvisioningGroup by @helen-frank in #177
• chore: release v0.1.4 by @jwcesign in #179
• fix: fix make update error by @jxs1211 in #181
• ci: push images to acr by @jwcesign in #184
• chore: change image to acr by @jwcesign in #186
• chore: print error beautiful by @jwcesign in #187
• fix: set default disk categories and size by @jwcesign in #188
• Feat/systemdisk size by @jxs1211 in #183
• refactor parsing userData by @jxs1211 in #190
• chore: delete redundant function parameters by @jwcesign in #191
• chore: upgrade the dependencies by @jwcesign in #185
• chore: cleanup duplicated nodeclass status by @jwcesign in #192
• chore: auto generate nodepool and nodeclaim crds by @jwcesign in #193
• doc: nominate helen-frank/daimaxiaxie as approver by @jwcesign in #194
• doc: typo by @jwcesign in #195
• feat: support data disk by @daimaxiaxie in #189
• fix: allow ClosedWithStock capacity type by @jwcesign in #196
• feat: update unavailableOfferings when adding a launchInstance with no available resources by @helen-frank in #197
• fix: fix nil pointer of disk conversation by @jwcesign in #198
• fix: fix release charts by @jwcesign in #201
• chore: release chart v0.1.5 by @jwcesign in #202
• fix: resolve spot capacity correctly by @jwcesign in #203
• fix: remove the unused env and add the required roles by @Vacant2333 in #207
• fix: consider vs zones when creating types offerings by @jwcesign in #205
• fix: select nodepool id when describe cluster attach scripts by @Vacant2333 in #208
• doc: add notice file for restrictions by @jwcesign in #211
• feat: support custom userData by @Vacant2333 in #212
• fix: replace the attach-script's runtime configurations by @Vacant2333 in #213
• fix: fix volumesize modification by @jwcesign in #216
• support set password and key pair in ecs node class by @kaiiyvwu in #218
• doc: add adopters by @jwcesign in #214

New Contributors

• @Aleegra made their first contribution in #171
• @kaiiyvwu made their first contribution in #218

Full Changelog: v0.1.8...v0.2.0

v0.1.8 release

What's Changed

• fix: replace the attach script runtime configuration by @Vacant2333 in #210

Full Changelog: v0.1.7...v0.1.8

v0.1.7 release

What's Changed

• fix: consider vs zones when creating types offerings by @jwcesign in #206
• fix: unused env and select nodepool ID by @Vacant2333 in #209

Full Changelog: v0.1.6...v0.1.7

v0.1.6 release

What's Changed

• Fix: Resolve spot capacity correctly. Sometimes, on-demand resources are out of capacity, but spot instances still have available capacity by @jwcesign in #204

Full Changelog: v0.1.5...v0.1.6

v0.1.5 release

What's Changed

• feat: update unavailableOfferings when adding a launchInstance with no available resources by @helen-frank in #199
• fix: allow ClosedWithStock capacity type by @jwcesign in #200

Full Changelog: v0.1.4...v0.1.5

v0.1.4 release

What's Changed

🛠️ Bug Fixes

• Fix: Do not throw an error when some instance types are unavailable for certain vSwitches. @jwcesign in #178

Full Changelog: v0.1.3...v0.1.4

v0.1.3 release

What's Changed

• fix: get image status even it's expired by @gitautomator in #170
• fix: fallback to OnDemand when Spot is not available by @gitautomator in #173

Full Changelog: v0.1.2...v0.1.3

v0.1.2 release

What's Changed

• chore: use SecurityGroupIds instad of SecurityGroupId for ProvisionGroup by @gitautomator in #163
• fix: calculate the overhead correctly by @gitautomator in #165
• chore: add tag to apg to enhance the SLA by @gitautomator in #164

Full Changelog: v0.1.1...v0.1.2

v0.1.1 release

What's Changed

• Cherry pick PR(153)/fix: give higher memory overhead percent by @gitautomator in #154

New Contributors

• @gitautomator made their first contribution in #154

Full Changelog: v0.1.0...v0.1.1

v0.1.0 release

Overview

🚀 We are excited to announce the release of the Karpenter Alibaba Cloud Provider! This new open-source provider extends Karpenter’s capabilities to Alibaba Cloud, enabling efficient, flexible, and cost-optimized node provisioning for Kubernetes clusters running on Alibaba Cloud infrastructure.

Key Features

1. Seamless Alibaba Cloud Integration:

   • Automatically provisions ECS instances for Kubernetes clusters on Alibaba Cloud.
   • Supports a wide range of ECS instance types, including general-purpose, compute-optimized, and GPU instances.

2. Dynamic Scaling:

   • Leverages Karpenter’s intelligent scheduling to dynamically adjust cluster capacity based on workload demands.
   • Ensures fast, on-demand scaling with minimal latency.

3. Cost Optimization:

   • Optimizes instance selection to reduce costs by leveraging preemptible instances (Spot Instances) when appropriate.
   • Balances cost and performance to meet workload requirements efficiently.

4. Custom Node Configuration:

   • Supports flexible configurations, including custom labels, taints, and annotations for node management.
   • Allows fine-tuned control over instance lifecycle and allocation.

5. High Availability:

   • Ensures robust provisioning with fault-tolerant instance placement strategies.

6. Open Source & Community-Driven:

   • Fully open-sourced to foster collaboration and continuous improvement.
   • Contributions and feedback from the community are highly encouraged.

Getting Started

Please check the doc.

Feedback and Contributions

We welcome your feedback and contributions to make this provider even better. Please visit the GitHub repository at GitHub Link to get involved.