The report suggests that many organizations are not keeping pace with updates as versions of Kubernetes are released three times a year. The challenge, of course, is not just rolling out updates to the core platform but also making sure applications don’t break as application programming interfaces (APIs) are either added or deprecated from the core platform.
In general, the report finds nearly half of organizations that deployed containerized applications are employing Kubernetes. More than 40% of those clusters, however, have lax privileges that pose a significant cybersecurity concern, the report finds.
The report also notes that just over 30% are also using more than 1,000 or more hosts in multiple clouds. However, only about 69% of hosts running the containerd runtime are using version 1.5 or 1.6, which are the actively supported versions. About 31% of containerd hosts are using versions 1.4 or older, which have passed their end-of-life support dates.
In addition, the report also finds well over a third (36%) of Datadog customers are now also using serverless container technologies.
More than 35% are now also using the ingress capability first made available in Kubernetes 1.19 to balance workloads, according to the report. Finally, Datadog noted it is seeing customers adopting both the Istio and Linkerd service mesh, with Istio being three times as popular.
Overall, the Datadog report finds the most widely used software running in containers managed via the Datadog platform are:
1. NGINX proxy software
2. Redis key/value store
3. Postgres database
4. Elasticsearch search engine
5. Kafka messaging software
6. RabbitMQ messaging software
7. MongoDB database
8. MySQL database
9. Calico network virtualization software
10. GitLab continuous integration/continuous delivery (CI/CD) platform
11: Vault secrets management software.
As fleets of Kubernetes clusters become more widely deployed in production environments, it’s clear that IT management issues are becoming more challenging.
John Kendall, senior product manager of containers at Datadog, says as more organizations adopt Kubernetes at scale there will be more reliance on managed Kubernetes services that provide organizations access to curated instances of the platform. Many of those service providers will continue to support older versions of Kubernetes beyond the official end-of-life of a release.
Regardless of how Kubernetes is managed, the stack of software deployed on top of these clusters is also getting larger. IT organizations not only need to find ways to manage Kubernetes clusters but also all the elements of an increasingly complex cloud-native software stack deployed on those clusters. The management of containers and the underlying Kubernetes clusters themselves are now just table stakes as cloud-native application environments are being deployed in mainstream IT environments, noted Kendall.
In the meantime, IT teams will need to decide whether to standardize on not just a version of Kubernetes but also a specific distribution after years of allowing individual development teams to roll out clusters whenever and however they saw fit.