This is part 2 of the updates on KubeCon, you can find the part 1 here
Day zero & one of the KubeCon US 2019 was filled with some interesting updates and sessions. We continued our conversations with our customers/partners and at the various booths with exciting products and solutions to new and emerging problems. Here is our view of the last two days of the conference.
Clusters are the new Cattle!
As organizations are managing more than just one cluster, the patterns of “multi-cluster” and emerging. One such Tweet from a talk I found emphasizes the fact that cluster should be things that can be automatically reproduced easily. Also, note the OPA for validation and mention of multi-cluster ingress – another pattern for operating multiple clusters.
— Kevin Stewart (@kstewart) November 21, 2019
Pack Cluster+Apps for Air-Gapped Environment
One of the interesting projects that I have been following for a while was Gravity. It solves the problem of deploying a cluster and all that is running in the cluster in an air-gapped environment in a unique way. The company announced that it has closed series A and it will be interesting to see the space.
Flux & Argo CD Join hands for GitOps
GitOps has become quite popular as a way to do CD, especially in the Kubernetes world. Two of the popular projects in the community were Flux and Argo CD. The two projects are going to collaborate and work on a joint solution so that end users can work with a single toolchain. This is a great and exciting development for the community and users.
MicroVMs – exciting future
One of the most exciting projects which are not mainstream Kubernetes/CNCF projects for me was the AWS Firecracker and a layer on top of that – Ignite. It would be interesting to see the use cases & development of managing MicroVMs using Kubernetes. On that note – I also maintain a repo for all resources related to MicroVMs, send a PR if you find anything interesting.
— Chris Hein (@christopherhein) November 21, 2019
Logs & observability – new perspectives
Anyone who has operated non-trivial clusters knows how hard the whole monitoring, logging, and observability can be. People have stories about how hard it is to manage an ElasticSearch cluster at scale. It is also why you see so many new companies that offer managed log management as a service. One interesting development in this space is a new log platform called Loki – which aims for “operational simplicity” has hit 1.0 milestone. The company behind Loki has been running the Loki in production for some time now and that gives them confidence in the stability. It will be interesting to see the convergence of monitoring, logging and tracing in the coming year!
KubeCon Videos are ready
For the sessions that you missed out or if you did not make it to KubeCon, the videos are already out there. I am going to watch some of them in transit on way back home 🙂 I am sure there are many more things that I missed in the conference – hopefully, I can do another post after watching some sessions that wanted to checkout!