I created an outline from the video The Best DevOps Tools, Frameworks, And Platforms In 2022, by Viktor Farcic at DevOps Toolkit. Subscribe and support him!
Category winners have links and more details.
Self Service (eg FAAS for local dev)
- convenience wrappers: ketch, shipa
- Knative: WINNER
- serverless solutions
- KubeVela: WINNER
- custom resource definitions
- “Open Application Model”, helping shift infrastructure left for developers.
Managed Containers
- fly.io: simple, limited
- GCR (Google Cloud Run): WINNER
- managed serverless workloads
- based on Knative
- ACI (Azure Container Images): not so good
- ACA (Azure Container Apps): new, good!
- AWS Lightsail
- AWS ECS: fargate/no fargate. Deprecate. Kubernetes competition
- AWS Lambda: can now use containers
Infrastructure and Services
- Chef and Puppet deprecated
- Ansible: on prem infrastucture
- Terraform: just works, supported for everybody
- Pulumi: Imperative (vs Declarative)
- eksctl.io: good but limited to kub clusters, only works aws eks
- Crossplane: WINNER
- new
- Works on entire CNCF ecosystem
- Drift detection
- Works with various GitOps tools like: Argo CD, flux, Rancher Fleet
- Supports all cloud K8s services
GitOps
- Flux: WINNER
- they invented GitOps.
- Argo CD: WINNER
- open source.
- Rancher Fleet: new, designed to run at scale.
Progressive Delivery
- Argo Rollouts: works well with ArgoCD
- Flagger: works well with Flux
Security
- snyk: comprehensive.
- Gatekeeper: powerful, difficult. OPA
- Kyverno: WINNER
- kube native competitor with Gatekeeper.
- Policy mgt.
- Kubescape: WINNER
- new, great promise
- Scans multi-cloud
- Combines in one “pane of glass”.
Development
CI/CD Pipelines
- Jenkins: Jenkins X?
- Circle CI: outlived Travis
- CodeFresh: new. Good for Argo workflows as a service
- Argo Workflows + Argo Events: WINNER
- not user friendly yet
- But best Kube native pipeline
- Tekton: like ArgoWF, less friendly than ArgoWF
- GitHub Actions: WINNER
- so easy but not kube native
- DevTron: opinionated tools. Good but new and rough.
Logging, Monitoring, Visualization And Troubleshooting
-
Logging
- Grafana Loki: WINNER
- self hosted option
- Grafana Loki: WINNER
-
Monitoring
- Prometheus: WINNER
-
Visualization
- Grafana Dashboards: WINNER
-
Troubleshooting
- Troubleshoot: WINNER
- troubleshoot as service to other companies/teams
- kubectl plugin
- Komodor: WINNER
- k8s native troubleshooting
- troubleshoot clusters you own. Gives much context.
- tracks changes, “ripples” across stack.
- Troubleshoot: WINNER
Kubernetes Dashboards
-
K9s: WINNER
- works in terminal like VIM
-
Kubeapps: collection of Helm charts
-
Kubesphere: opinionated
-
K8slens: polished
-
Octant: dont use
-
Rancher Kubernetes Dashboard: WINNER
- was dont use 2021, but better now
- Multi clusters
- Web UI
Containers: always run in K8s, never Docker
- Docker: deprecated. bloated. no compose, no swarm. no K8s.
- BuildX: demon, ok for local. Already integrated with Docker. No good for pipeline because is demon.
- Kaniko: WINNER for prod
- remote builds
- “One shot” execution, no demon
- Shipwright: WINNER for prod, using Kaniko
- CDF member
- Engine wrapper
- Engine agnostic (use Kaniko)
- NerdCTL: WINNER local builds
- From ContainerD
- Not for production, good for laptop
Kubernetes
Managed Clusters (as a service)
Big Vendors
- Amazon EKS: lags others
- AWS Fargate: strange beast
- Azure AKS: good but flaky
- GKE (Google Kube Engine): WINNER
- always ahead
- Autopilot: fulllly managed GKE, expensive
Small Vendors
- Linode LKE: not HA. Cheap.
- Digital Ocean DOKs: not HA. Cheap.
- Civo: WINNER
- Not HA
- Cheap
- K3s so best of small guys
Self Serve
- OpenShift
- Rancher