Read on Twitter
Thread: So it is #KubeCon time, which of course means lots of interesting data from various vendors. The always interesting @datadoghq report has been published https://www.datadoghq.com/container-report/ /1
Lots of things here match into what we encounter on customer inquiries. Lets call out three big @kubernetesio trends before we look at the DataDog report
* Kubernetes on cloud is way, way bigger in terms of number of clusters than all of the other vendors combined #KubeCon /2
* Kubernetes on cloud is way, way bigger in terms of number of clusters than all of the other vendors combined #KubeCon /2
* Self managed and distro based #k8s is decreasing in relative market share to the provider managed offerings. The recent OpenShift price drops and the bottom tier VMware Tanzu sku show the pressures here - you need a value add beyond just #k8s. #KubeCon /3
* On prem installs have larger & more nodes, less churn, less experiments and more production. There is lots of churn in the cloud providers. That said you are unlikely to be using DataDog to monitor your experimental #k8s cluster. #KubeCon /4
And onto the report. Firstly @kubernetesio is far from the only game in town. It is growing, but for @datadoghq its still just over half of the containers they observe.
Put another way it is (correctly) very far from a done deal that #k8s is the only way. #KubeCon /5
Put another way it is (correctly) very far from a done deal that #k8s is the only way. #KubeCon /5
The managed @kubernetesio offerings dominate on @Azure @awscloud and @googlecloud. I only see this accelerating in my discussions, particuarly away from self managed upstream #k8s on #iaas. /cc @datadoghq #KubeCon /6
Lots of growth in @awscloud Fargate. Again no surprises here, once clients master the initial aspects of using #k8s, cutting costs is the next step /cc @datadoghq #kubecon /7
On the other side of the Fargate usage, badly sized nodes is really common. Old school capacity planning has not caught up with new tech in lots of orgs. Availability is factor here - many of these regions came up in client calls prior to Oct https://aws.amazon.com/about-aws/whats-new/2020/10/amazon-eks-adds-fargate-support-10-additional-aws-regions/ #KubeCon /8
The #servicemesh figures are welcome to see. The majority of inquiries I have in this space is a kicking the tires inquiry over production usage. It is indeed very early days. #KubeCon /9
It is notable how much @kubernetesio is growing up - as in people are running lots of versions a year or more old. Classic sustaining problem and why people pay vendors for LTS. #KubeCon /10
* read the @datadoghq notes on the percentages
* read the @datadoghq notes on the percentages
And finally the container registry data. The growth of cloud provided registries makes sense. And it is only going to continue growing. It will be interesting to see if the @Docker rate limits have an impact longer term here. /cc @datadoghq #KubeCon /11
Also the need for validation and verification in the software supply chain is going to add lots more pressures here. Lots of opportunities for various vendors to do the heavy lifting. This is also why I expect to see @Paketo_io & @buildpacks_io in general become v. important /12
Some other lovely nuggets in the report. I am always surprised, then happy as its almost always for the right reasons, to see how much #Redis is in use for instance. /13
