Author: Bobby Salamat

Kubernetes is well-known for running scalable workloads. It scales your workloads based on their resource usage. When a workload is scaled up, more instances of the application get created. When the application is critical for your product, you want to make sure that these new instances are scheduled even when your cluster is under resource pressure. One obvious solution to this problem is to over-provision your cluster resources to have some amount of slack resources available for scale-up situations. This approach often works, but costs more as you would have to pay for the resources that are idle most of the time.

Pod priority and preemption is a scheduler feature made generally available in Kubernetes 1.14 that allows you to achieve high levels of scheduling confidence for your critical workloads without overprovisioning your clusters. It also provides a way to improve resource utilization in your clusters without sacrificing the reliability of your essential workloads.

Guaranteed scheduling with controlled cost

Kubernetes Cluster Autoscaler is an excellent tool in the ecosystem which adds more nodes to your cluster when your applications need them. However, cluster autoscaler has some limitations and may not work for all users:

• It does not work in physical clusters.
• Adding more nodes to the cluster costs more.
• Adding nodes is not instantaneous and could take minutes before those nodes become available for scheduling.

An alternative is Pod Priority and Preemption. In this approach, you combine multiple workloads in a single cluster. For example, you may run your CI/CD pipeline, ML workloads, and your critical service in the same cluster. When multiple workloads run in the same cluster, the size of your cluster is larger than a cluster that you would use to run only your critical service. If you give your critical service the highest priority and your CI/CD and ML workloads lower priority, when your service needs more computing resources, the scheduler preempts (evicts) enough pods of your lower priority workloads, e.g., ML workload, to allow all your higher priority pods to schedule.

With pod priority and preemption you can set a maximum size for your cluster in the Autoscaler configuration to ensure your costs get controlled without sacrificing availability of your service. Moreover, preemption is much faster than adding new nodes to the cluster. Within seconds your high priority pods are scheduled, which is critical for latency sensitive services.

Improve cluster resource utilization

Cluster operators who run critical services learn over time a rough estimate of the number of nodes that they need in their clusters to achieve high service availability. The estimate is usually conservative. Such estimates take bursts of traffic into account to find the number of required nodes. Cluster autoscaler can be configured never to reduce the size of the cluster below this level. The only problem is that such estimates are often conservative and cluster resources may remain underutilized most of the time. Pod priority and preemption allows you to improve resource utilization significantly by running a non-critical workload in the cluster.

The non-critical workload may have many more pods that can fit in the cluster. If you give a negative priority to your non-critical workload, Cluster Autoscaler does not add more nodes to your cluster when the non-critical pods are pending. Therefore, you won’t incur higher expenses. When your critical workload requires more computing resources, the scheduler preempts non-critical pods and schedules critical ones.

The non-critical pods fill the “holes” in your cluster resources which improves resource utilization without raising your costs.

Get Involved

If you have feedback for this feature or are interested in getting involved with the design and development, join the Scheduling Special Interest Group.

Authors: Andrew Sy Kim (VMware), Mike Crute (AWS), Walter Fender (Google)

Approximately 9 months ago, the Kubernetes community agreed to form the Cloud Provider Special Interest Group (SIG). The justification was to have a single governing SIG to own and shape the integration points between Kubernetes and the many cloud providers it supported. A lot has been in motion since then and we’re here to share with you what has been accomplished so far and what we hope to see in the future.

The Mission

First and foremost, I want to share what the mission of the SIG is, because we use it to guide our present & future work. Taken straight from our charter, the mission of the SIG is to simplify, develop and maintain cloud provider integrations as extensions, or add-ons, to Kubernetes clusters. The motivation behind this is two-fold: to ensure Kubernetes remains extensible and cloud agnostic.

The Current State of Cloud Providers

In order to gain a forward looking perspective to our work, I think it’s important to take a step back to look at the current state of cloud providers. Today, each core Kubernetes component (except the scheduler and kube-proxy) has a –cloud-provider flag you can configure to enable a set of functionalities that integrate with the underlying infrastructure provider, a.k.a the cloud provider. Enabling this integration unlocks a wide set of features for your clusters such as: node address & zone discovery, cloud load balancers for Services with Type=LoadBalancer, IP address management, and cluster networking via VPC routing tables. Today, the cloud provider integrations can be done either in-tree or out-of-tree.

In-Tree & Out-of-Tree Providers

In-tree cloud providers are the providers we develop & release in the main Kubernetes repository. This results in embedding the knowledge and context of each cloud provider into most of the Kubernetes components. This enables more native integrations such as the kubelet requesting information about itself via a metadata service from the cloud provider.

Out-of-tree cloud providers are providers that can be developed, built, and released independent of Kubernetes core. This requires deploying a new component called the cloud-controller-manager which is responsible for running all the cloud specific controllers that were previously run in the kube-controller-manager.

When cloud provider integrations were initially developed, they were developed natively (in-tree). We integrated each provider close to the core of Kubernetes and within the monolithic repository that is k8s.io/kubernetes today. As Kubernetes became more ubiquitous and more infrastructure providers wanted to support Kubernetes natively, we realized that this model was not going to scale. Each provider brings along a large set of dependencies which increases potential vulnerabilities in our code base and significantly increases the binary size of each component. In addition to this, more of the Kubernetes release notes started to focus on provider specific changes rather than core changes that impacted all Kubernetes users.

In late 2017, we developed a way for cloud providers to build integrations without adding them to the main Kubernetes tree (out-of-tree). This became the de-facto way for new infrastructure providers in the ecosystem to integrate with Kubernetes. Since then, we’ve been actively working towards migrating all cloud providers to use the out-of-tree architecture as most clusters today are still using the in-tree cloud providers.

Looking Ahead

Looking ahead, the goal of the SIG is to remove all existing in-tree cloud providers in favor of their out-of-tree equivalents with minimal impact to users. In addition to the core cloud provider integration mentioned above, there are more extension points for cloud integrations like CSI and the image credential provider that are actively being worked on for v1.15. Getting to this point would mean that Kubernetes is truly cloud-agnostic with no native integrations for any cloud provider. By doing this work we empower each cloud provider to develop and release new versions at their own cadence independent of Kubernetes. We’ve learned by now that this is a large feat with a unique set of challenges. Migrating workloads is never easy, especially when it’s an essential part of the control plane. Providing a safe and easy migration path between in-tree and out-of-tree cloud providers is of the highest priority for our SIG in the upcoming releases. If any of this sounds interesting to you, I encourage you to check out of some of our KEPs and get in touch with our SIG by joining the mailing list or our slack channel (#sig-cloud-provider in Kubernetes slack).

Authors:
Paris Pittman (Google), Jonas Rosland (VMware)

tl;dr – click here for Barcelona Contributor Summit information.

As our contributing community grows in great numbers, with more than 16,000 contributors this year across 150+ GitHub repositories, it’s important to provide face to face connections for our large distributed teams to have opportunities for collaboration and learning. In Contributor Experience, our methodology with planning events is a lot like our documentation; we build from personas – interests, skills, and motivators to name a few. This way we ensure there is valuable content and learning for everyone.

We build the contributor summits around you:

• New Contributor
• Current Contributor
  • docs
  • code
  • community management
• Subproject OWNERs
• Special Interest Group (SIG) / Working Group (WG) Chair or Tech Lead
• Active Contributors
• Casual Contributors

These personas combined with ample feedback from previous events, produce the altogether experience that welcomed over 600 contributors in Copenhagen (May), Shanghai(November), and Seattle(December) in 2018. Seattle’s event drew over 300+ contributors, equal to Shanghai and Copenhagen combined, for the 6th contributor event in Kubernetes history. In true Kubernetes fashion, we expect another record breaking year of attendance. We’ve pre-ordered 900+ contributor patches, a tradition, and we are looking forward to giving them to you!

With that said…
Save the Dates:
Barcelona: May 19th (evening) and 20th (all day)
Shanghai: June 24th (all day)
San Diego: November 18th, 19th, and activities in KubeCon/CloudNativeCon week

In an effort of continual improvement, here’s what to expect from us this year:

• Large new contributor workshops and contributor socials at all three events expected to break previous attendance records
• A multiple track event in San Diego for all contributor types including workshops, birds of a feather, lightning talks and more
• Addition of a “201” / “Intermediate” edition of the new contributor workshop in San Diego
• An event website!
• Follow along with updates: kubernetes-dev@googlegroups.com is our main communication hub as always; however, we will also blog here, our Thursday Kubernetes Community Meeting, twitter, SIG meetings, event site, discuss.kubernetes.io, and #contributor-summit on Slack.
• Opportunities to get involved: We still have 2019 roles available!
  Reach out to Contributor Experience via community@kubernetes.io, stop by a Wednesday SIG update meeting, or catch us on Slack (#sig-contribex).

Thanks!

Our 2018 crew ?
Jorge Castro, Paris Pittman, Bob Killen, Jeff Sica, Megan Lehn, Guinevere Saenger, Josh Berkus, Noah Abrahams, Yang Li, Xiangpeng Zhao, Puja Abbassi, Lindsey Tulloch, Zach Corleissen, Tim Pepper, Ihor Dvoretskyi, Nancy Mohamed, Chris Short, Mario Loria, Jason DeTiberus, Sahdev Zala, Mithra Raja

And an introduction to our 2019 crew (a thanks in advance 😉 )…
Jonas Rosland, Josh Berkus, Paris Pittman, Jorge Castro, Bob Killen, Deb Giles, Guinevere Saenger, Noah Abrahams, Yang Li, Xiangpeng Zhao, Puja Abbassi, Rui Chen, Tim Pepper, Ihor Dvoretskyi, Dawn Foster

Relive Seattle Contributor Summit

? 80% growth rate since the Austin 2017 December event
? Event waiting list: 103
? 76 contributors were on-boarded through the New Contributor Workshop
? 92% of the current contributors RSVPs attended and of those:
??‍? 25% were Special Interest Group or Working Group Chairs or Tech Leads
? 70% were eligible to vote in the last steering committee election
? 20+ Sessions
? Most watched to date: Technical Vision, Security, API Code Base Tour
? Top 3 according to survey: Live API Code Review, Deflaking Unconference, Technical Vision
? ? 160 attendees for the social at Garage on Sunday night where we sunk eight balls and recorded strikes (out in some cases)
? Special recognition: SIG Storage, @dims, and @jordan
? Pictures (special thanks to rdodev)
Garage Pic
Reg Desk

“I love Contrib Summit! The intros and deep dives during KubeCon were a great extension of Contrib Summit. Y’all did an excellent job in the morning to level set expectations and prime everyone.” – julianv
“great work! really useful and fun!” – coffeepac

Authors:
Paris Pittman (Google), Jonas Rosland (VMware)

tl;dr – click here for Barcelona Contributor Summit information.

As our contributing community grows in great numbers, with more than 16,000 contributors this year across 150+ GitHub repositories, it’s important to provide face to face connections for our large distributed teams to have opportunities for collaboration and learning. In Contributor Experience, our methodology with planning events is a lot like our documentation; we build from personas – interests, skills, and motivators to name a few. This way we ensure there is valuable content and learning for everyone.

We build the contributor summits around you:

• New Contributor
• Current Contributor
  • docs
  • code
  • community management
• Subproject OWNERs
• Special Interest Group (SIG) / Working Group (WG) Chair or Tech Lead
• Active Contributors
• Casual Contributors

These personas combined with ample feedback from previous events, produce the altogether experience that welcomed over 600 contributors in Copenhagen (May), Shanghai(November), and Seattle(December) in 2018. Seattle’s event drew over 300+ contributors, equal to Shanghai and Copenhagen combined, for the 6th contributor event in Kubernetes history. In true Kubernetes fashion, we expect another record breaking year of attendance. We’ve pre-ordered 900+ contributor patches, a tradition, and we are looking forward to giving them to you!

With that said…
Save the Dates:
Barcelona: May 19th (evening) and 20th (all day)
Shanghai: June 24th (all day)
San Diego: November 18th, 19th, and activities in KubeCon/CloudNativeCon week

In an effort of continual improvement, here’s what to expect from us this year:

• Large new contributor workshops and contributor socials at all three events expected to break previous attendance records
• A multiple track event in San Diego for all contributor types including workshops, birds of a feather, lightning talks and more
• Addition of a “201” / “Intermediate” edition of the new contributor workshop in San Diego
• An event website!
• Follow along with updates: kubernetes-dev@googlegroups.com is our main communication hub as always; however, we will also blog here, our Thursday Kubernetes Community Meeting, twitter, SIG meetings, event site, discuss.kubernetes.io, and #contributor-summit on Slack.
• Opportunities to get involved: We still have 2019 roles available!
  Reach out to Contributor Experience via community@kubernetes.io, stop by a Wednesday SIG update meeting, or catch us on Slack (#sig-contribex).

Thanks!

Our 2018 crew ?
Jorge Castro, Paris Pittman, Bob Killen, Jeff Sica, Megan Lehn, Guinevere Saenger, Josh Berkus, Noah Abrahams, Yang Li, Xiangpeng Zhao, Puja Abbassi, Lindsey Tulloch, Zach Corleissen, Tim Pepper, Ihor Dvoretskyi, Nancy Mohamed, Chris Short, Mario Loria, Jason DeTiberus, Sahdev Zala, Mithra Raja

And an introduction to our 2019 crew (a thanks in advance 😉 )…
Jonas Rosland, Josh Berkus, Paris Pittman, Jorge Castro, Bob Killen, Deb Giles, Guinevere Saenger, Noah Abrahams, Yang Li, Xiangpeng Zhao, Puja Abbassi, Rui Chen, Tim Pepper, Ihor Dvoretskyi, Dawn Foster

Relive Seattle Contributor Summit

? 80% growth rate since the Austin 2017 December event
? Event waiting list: 103
? 76 contributors were on-boarded through the New Contributor Workshop
? 92% of the current contributors RSVPs attended and of those:
??‍? 25% were Special Interest Group or Working Group Chairs or Tech Leads
? 70% were eligible to vote in the last steering committee election
? 20+ Sessions
? Most watched to date: Technical Vision, Security, API Code Base Tour
? Top 3 according to survey: Live API Code Review, Deflaking Unconference, Technical Vision
? ? 160 attendees for the social at Garage on Sunday night where we sunk eight balls and recorded strikes (out in some cases)
? Special recognition: SIG Storage, @dims, and @jordan
? Pictures (special thanks to rdodev)
Garage Pic
Reg Desk

“I love Contrib Summit! The intros and deep dives during KubeCon were a great extension of Contrib Summit. Y’all did an excellent job in the morning to level set expectations and prime everyone.” – julianv
“great work! really useful and fun!” – coffeepac

Authors:
Paris Pittman (Google), Jonas Rosland (VMware)

tl;dr – click here for Barcelona Contributor Summit information.

As our contributing community grows in great numbers, with more than 16,000 contributors this year across 150+ GitHub repositories, it’s important to provide face to face connections for our large distributed teams to have opportunities for collaboration and learning. In Contributor Experience, our methodology with planning events is a lot like our documentation; we build from personas – interests, skills, and motivators to name a few. This way we ensure there is valuable content and learning for everyone.

We build the contributor summits around you:

• New Contributor
• Current Contributor
  • docs
  • code
  • community management
• Subproject OWNERs
• Special Interest Group (SIG) / Working Group (WG) Chair or Tech Lead
• Active Contributors
• Casual Contributors

These personas combined with ample feedback from previous events, produce the altogether experience that welcomed over 600 contributors in Copenhagen (May), Shanghai(November), and Seattle(December) in 2018. Seattle’s event drew over 300+ contributors, equal to Shanghai and Copenhagen combined, for the 6th contributor event in Kubernetes history. In true Kubernetes fashion, we expect another record breaking year of attendance. We’ve pre-ordered 900+ contributor patches, a tradition, and we are looking forward to giving them to you!

With that said…
Save the Dates:
Barcelona: May 19th (evening) and 20th (all day)
Shanghai: June 24th (all day)
San Diego: November 18th, 19th, and activities in KubeCon/CloudNativeCon week

In an effort of continual improvement, here’s what to expect from us this year:

• Large new contributor workshops and contributor socials at all three events expected to break previous attendance records
• A multiple track event in San Diego for all contributor types including workshops, birds of a feather, lightning talks and more
• Addition of a “201” / “Intermediate” edition of the new contributor workshop in San Diego
• An event website!
• Follow along with updates: kubernetes-dev@googlegroups.com is our main communication hub as always; however, we will also blog here, our Thursday Kubernetes Community Meeting, twitter, SIG meetings, event site, discuss.kubernetes.io, and #contributor-summit on Slack.
• Opportunities to get involved: We still have 2019 roles available!
  Reach out to Contributor Experience via community@kubernetes.io, stop by a Wednesday SIG update meeting, or catch us on Slack (#sig-contribex).

Thanks!

Our 2018 crew ?
Jorge Castro, Paris Pittman, Bob Killen, Jeff Sica, Megan Lehn, Guinevere Saenger, Josh Berkus, Noah Abrahams, Yang Li, Xiangpeng Zhao, Puja Abbassi, Lindsey Tulloch, Zach Corleissen, Tim Pepper, Ihor Dvoretskyi, Nancy Mohamed, Chris Short, Mario Loria, Jason DeTiberus, Sahdev Zala, Mithra Raja

And an introduction to our 2019 crew (a thanks in advance 😉 )…
Jonas Rosland, Josh Berkus, Paris Pittman, Jorge Castro, Bob Killen, Deb Giles, Guinevere Saenger, Noah Abrahams, Yang Li, Xiangpeng Zhao, Puja Abbassi, Rui Chen, Tim Pepper, Ihor Dvoretskyi, Dawn Foster

Relive Seattle Contributor Summit

? 80% growth rate since the Austin 2017 December event
? Event waiting list: 103
? 76 contributors were on-boarded through the New Contributor Workshop
? 92% of the current contributors RSVPs attended and of those:
??‍? 25% were Special Interest Group or Working Group Chairs or Tech Leads
? 70% were eligible to vote in the last steering committee election
? 20+ Sessions
? Most watched to date: Technical Vision, Security, API Code Base Tour
? Top 3 according to survey: Live API Code Review, Deflaking Unconference, Technical Vision
? ? 160 attendees for the social at Garage on Sunday night where we sunk eight balls and recorded strikes (out in some cases)
? Special recognition: SIG Storage, @dims, and @jordan
? Pictures (special thanks to rdodev)
Garage Pic
Reg Desk

“I love Contrib Summit! The intros and deep dives during KubeCon were a great extension of Contrib Summit. Y’all did an excellent job in the morning to level set expectations and prime everyone.” – julianv
“great work! really useful and fun!” – coffeepac

Authors:
Paris Pittman (Google), Jonas Rosland (VMware)

tl;dr – click here for Barcelona Contributor Summit information.

As our contributing community grows in great numbers, with more than 16,000 contributors this year across 150+ GitHub repositories, it’s important to provide face to face connections for our large distributed teams to have opportunities for collaboration and learning. In Contributor Experience, our methodology with planning events is a lot like our documentation; we build from personas – interests, skills, and motivators to name a few. This way we ensure there is valuable content and learning for everyone.

We build the contributor summits around you:

• New Contributor
• Current Contributor
  • docs
  • code
  • community management
• Subproject OWNERs
• Special Interest Group (SIG) / Working Group (WG) Chair or Tech Lead
• Active Contributors
• Casual Contributors

These personas combined with ample feedback from previous events, produce the altogether experience that welcomed over 600 contributors in Copenhagen (May), Shanghai(November), and Seattle(December) in 2018. Seattle’s event drew over 300+ contributors, equal to Shanghai and Copenhagen combined, for the 6th contributor event in Kubernetes history. In true Kubernetes fashion, we expect another record breaking year of attendance. We’ve pre-ordered 900+ contributor patches, a tradition, and we are looking forward to giving them to you!

With that said…
Save the Dates:
Barcelona: May 19th (evening) and 20th (all day)
Shanghai: June 24th (all day)
San Diego: November 18th, 19th, and activities in KubeCon/CloudNativeCon week

In an effort of continual improvement, here’s what to expect from us this year:

• Large new contributor workshops and contributor socials at all three events expected to break previous attendance records
• A multiple track event in San Diego for all contributor types including workshops, birds of a feather, lightning talks and more
• Addition of a “201” / “Intermediate” edition of the new contributor workshop in San Diego
• An event website!
• Follow along with updates: kubernetes-dev@googlegroups.com is our main communication hub as always; however, we will also blog here, our Thursday Kubernetes Community Meeting, twitter, SIG meetings, event site, discuss.kubernetes.io, and #contributor-summit on Slack.
• Opportunities to get involved: We still have 2019 roles available!
  Reach out to Contributor Experience via community@kubernetes.io, stop by a Wednesday SIG update meeting, or catch us on Slack (#sig-contribex).

Thanks!

Our 2018 crew ?
Jorge Castro, Paris Pittman, Bob Killen, Jeff Sica, Megan Lehn, Guinevere Saenger, Josh Berkus, Noah Abrahams, Yang Li, Xiangpeng Zhao, Puja Abbassi, Lindsey Tulloch, Zach Corleissen, Tim Pepper, Ihor Dvoretskyi, Nancy Mohamed, Chris Short, Mario Loria, Jason DeTiberus, Sahdev Zala, Mithra Raja

And an introduction to our 2019 crew (a thanks in advance 😉 )…
Jonas Rosland, Josh Berkus, Paris Pittman, Jorge Castro, Bob Killen, Deb Giles, Guinevere Saenger, Noah Abrahams, Yang Li, Xiangpeng Zhao, Puja Abbassi, Rui Chen, Tim Pepper, Ihor Dvoretskyi, Dawn Foster

Relive Seattle Contributor Summit

? 80% growth rate since the Austin 2017 December event
? Event waiting list: 103
? 76 contributors were on-boarded through the New Contributor Workshop
? 92% of the current contributors RSVPs attended and of those:
??‍? 25% were Special Interest Group or Working Group Chairs or Tech Leads
? 70% were eligible to vote in the last steering committee election
? 20+ Sessions
? Most watched to date: Technical Vision, Security, API Code Base Tour
? Top 3 according to survey: Live API Code Review, Deflaking Unconference, Technical Vision
? ? 160 attendees for the social at Garage on Sunday night where we sunk eight balls and recorded strikes (out in some cases)
? Special recognition: SIG Storage, @dims, and @jordan
? Pictures (special thanks to rdodev)
Garage Pic
Reg Desk

“I love Contrib Summit! The intros and deep dives during KubeCon were a great extension of Contrib Summit. Y’all did an excellent job in the morning to level set expectations and prime everyone.” – julianv
“great work! really useful and fun!” – coffeepac

Artificial intelligence has the potential to change lots of things (in good ways) for humanity. Some people, notably Elon Musk, fear an AI arms race could lead to another world war. AI has shown that it is capable of beating humans in games that humans have mastered over many years, such as Google’s AlphaGo AI defeating Go world champion Lee

Few things are more annoying than when a Windows update wreaks havoc on your system. It’s a case of the cure being worse than the disease, as Ray Stevens sang about in his comical song “Doctor Doctor (Have Mercy On Me).” We saw this play out recently when it was reported that a Patch Tuesday roll out was causing boot problems on some systems

When most people think about the Ford Mustang, they imagine a rear-wheel drive beast powered by a thumping V8 engine. However, in recent years Ford has reintroduced a turbocharged four-cylinder to the lineup and it is by no means a slouch in the performance arena.
Today, Ford has introduced the new 2.3L High Performance Package for the four-banger

We know that Google is working on new smartphones for the Pixel family; after all, the company leaked specs on the Pixel 3a and 3a XL on its website. Details of those devices have also leaked that show the mid-range smartphones will cost around $500 with the Pixel 3a XL naturally being the more expensive version. Google is now teasing the