Hybrid Cloud - nothing to worry about! 🌥️
summary
Argues a hybrid approach: production on cheaper dedicated servers, bursting and staging on AWS, and cloud used for durable backups; running a Kubernetes cluster on bare metal, building live DB replication, and keeping ephemeral test envs in the cloud. Claims this setup saved 2–3 engineer salaries versus pure cloud, uses KVM remote access, LVM for easy disk expansion, and unattended installs for scale. Advises delaying VMs/containers until multiple teams need isolation; reports the system works with few infra outages and most incidents coming from application bugs.
AI-written summary · about
chapters (17)
- 0:00Intro
- 1:16Infrastructure
- 2:10Platform as a Service
- 4:20Server Shopping
- 4:56Virtual Machines
- 6:30Dedicated Servers
- 7:21Twisto
- 8:02Twisto infrastructure
- 11:20Cost
- 12:49Benefits
- 14:29Resolution of the question
- 15:39KVM
- 16:49Disks
- 17:37Virtualization
- 18:39Unattended installation
- 20:03Last tip
- 20:31Questions
related talks
26mCloud Agnostic Webscale
20mAre the days of on-premise servers over?
25mBreaking Free From the Cloud: Banking on Self-Hosted Kubernetes
43mBuilding Monetary Confidence in the Cloud
37mHybrid OnPrem + Cloud Architectures
18mHybrid cloud dos and don'ts
09mHybrid Cloud, The Best of Both Worlds
38m50 Shades of Multi-Cloud Infrastructure