Applications architecture spanning multiple clouds
summary
Projects start with containerizing apps and running OpenShift clusters per site, then solve CI/CD and replication pragmatically: initial Jenkins pipelines fail when one cluster rejects a deploy so teams move image distribution to Quay geo-replication and config management to Argo CD GitOps which enforces desired state and per-cluster overlays. Data sync uses either app-level Galera/JDG replication or object sync via NooBaa cloud object gateway. Monitoring tests show Grafana pulls can be ~two times cheaper than Prometheus federation. Service mesh and Submariner handle cross-cluster networking with a single control-plane preference.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 2:38Reasons for multi-cloud
- 6:11Multi-cloud deployment scenarios
- 7:15Containerization strategies
- 14:32CI/CD across clouds
- 24:02Networking challenges
- 30:01Data replication complexities
- 36:40Management tools and platforms
- 40:23Monitoring and cost
- 43:34Security policies
- 47:12Q&A
related talks
26mManaging GitOps Deployments in Multi-cluster Production Environments
09mInfra-like-apps - GitOpsifying Cloud Natively Managed Infrastructure with...
30mApp Delivery in Cloud Native: Where are We? What's Next?
34mInfrastructure at Scale Using ArgoCD
24mScaling GitOps Using ArgoCD Application-Sets: Lesson Learnt
27mGitOpsify Everything: When Crossplane Meets Argo CD
30mMigrating a monolith to Cloud-Native and the stumbling blocks
36mKEYNOTE: Taking a step back and leveraging GitOps to wrangle your clusters and projects