Fairwinds provides Managed Kubernetes-as-a-Service & open source software to enable organizations to run & optimize mission-critical Kubernetes infrastructure
USA
- Tweets 3,377
- Following 762
- Followers 1,391
- Likes 714
Meet Fairwinds, the Kubernetes people, at AWS re:Invent booth #1835. We’ll be there to hear how you’re managing the infrastructure running your AI workloads. Where does your team need support? See you in Las Vegas, November 30–December 4!
hubs.li/Q04BZhQV0
#AWSreInvent #K8s
59
For Noonlight, the availability bar is explicit: an AWS availability zone failure can't cause an outage, and a failed video or alarm creation through its APIs is unacceptable.
Read how the team approached kOps-to-EKS migration: bit.ly/4zVQjYs #Kubernetes #availability
65
You built and sold an AI platform. Pre-flight checks pass, but the customer's security review is still pending.
Is the installation ready to proceed?
Learn how technical checks and customer approvals fit into a deployment readiness gate: hubs.li/Q04zvpkY0 #kubernetes #ai
26
Kubernetes 1.37 extends in-place pod resizing with init container support in GA, making zero-downtime VPA more practical. Restart-free vertical scaling requires cgroup v2 on node pools as cgroup v1 defaults tighten.
hubs.li/Q04wMs3w0
#Kubernetes #DevOps #CloudNative
45
You built and sold an AI platform. The install call is booked. Your engineers are ready. The customer's security review is still pending.
What makes that install date credible?
See what a Definition of Ready should establish before implementation: hubs.li/Q04zvpfW0 #k8s
30
Our n8n chart migration on EKS ran into a gp3 storage constraint: the volume could not attach to multiple main pods across nodes. We treated the move from the community chart as a cutover.
See how we did it and get an example: hubs.li/Q04ydpql0 #n8n #kubernetes #eks
1
43
Fairwinds retweeted
The talks are great, but community is why people return to KubeCon + CloudNativeCon.
Zdeněk Váňa shares why connecting with peers is the best part.
Register: bit.ly/4hwqfuN
#CNCF #KubeCon
2
4
1,329
Your next release needs a newer Kubernetes version. Your customer can’t upgrade yet. How much notice did they get, and is there a documented workaround? Your compatibility policy shapes the conversation long before you get the support ticket.
hubs.li/Q04BjKPd0 #Kubernetes
32
Planning a Kubernetes 1.37 upgrade? Clear out any legacy v1alpha2 PodGroup resources from earlier gang-scheduling tests first to avoid API server validation errors, & verify cgroup v2 support across all active node pools.
Learn more about the release:
hubs.li/Q04wMs3W0 #K8s
35
Kubernetes rightsizing should reflect more than a single utilization snapshot. Review routine workload behavior, demand peaks, restarts, traffic patterns, and available capacity. Read about the data that informs request + limit changes. bit.ly/4rEjOdD #K8s #Observability
41
Noonlight's relationship with Fairwinds extends beyond day-to-day cluster operations. The team uses Fairwinds as a Slack sounding board for service choices, Terraform questions, and other infrastructure-adjacent decisions.
Who's your go-to for expertise?
bit.ly/4qQDSJk
34
Your customer bought the AI platform, but installation depends on access approvals and firewall changes from other teams. Who has committed to the install date? See how to account for those dependencies in your delivery plan:
hubs.li/Q04zvpkZ0 #kubernetes #aiplatforms
30
Kubernetes resource requests affect more than cost. When requests exceed actual workload needs, Pods can be harder to place and Nodes can be harder to consolidate. Read about how rightsizing affects cluster capacity.
bit.ly/4APaoR0
#Kubernetes #CloudOps
33
This production n8n deployment on EKS separated workflow execution from webhook traffic. Workers pulled jobs from Redis, public webhooks used an internet-facing load balancer, and internal webhooks used a private hostname.
See how we did it: hubs.li/Q04ydpsH0 #n8n #k8s
1
41
Gang scheduling reaches a major milestone in Kubernetes 1.37. The API split is complete with Workload and PodGroup moving to v1beta1, giving the scheduler native all-or-nothing placement controls to prevent AI batch job deadlocks.
Learn more:
hubs.li/Q04wMrSW0 #k8s #AI
40
Kubernetes rightsizing is more than a cost exercise. CPU and memory requests and limits shape workload placement, cluster capacity, application performance, and compute spend. Read about the data + decisions needed for reliable review.
bit.ly/3VdY3Wa
#Kubernetes #FinOps
1
72
Infrastructure projects compete directly with product work for many engineering teams. Noonlight partnered with Fairwinds to managed the Kubernetes layer, leaving engineers more time to build the life safety products and services customers rely on: bit.ly/4cun3y1 #K8s
33
Fairwinds CTO Andy Suderman ran this live at AWS Community Day Midwest: an LLM inference endpoint on EKS, deployment to a real response. Watch the replay. hubs.li/Q04qK83p0 #AWSCommunityDay #Kubernetes #EKS
33
What are the potential costs of getting Kubernetes operations wrong? Beyond cloud spend or an incident, they include the engineering time spent on upgrades, networking, monitoring, and infrastructure issues that compete with product work.
hubs.li/Q04vJCh80
#Kubernetes #SRE
41
Self-service Kubernetes often leads to dozens of clusters and one architect carrying most of the operational load. When do expert teams decide the platform layer should move to Managed Kubernetes as a Service?
hubs.li/Q04rdv0P0 #Kubernetes #ManagedKaaS
30