DR basics
Cloud DR basics: how cloud infrastructure comes back, lesson by lesson
Short answer
Nine lessons in three tracks: what must be in place before an incident, how the environment comes back during one, and what else has to come back besides servers and networks. Each lesson takes a few minutes and links to the sources it uses. Start with what goes missing when a region goes down if you are new to cloud DR.
What has to be in place before an incident?
The parts of a cloud DR plan that cannot be improvised on the day.
- Lesson 1 · 3 min readWhat Goes Missing When a Cloud Region Goes DownA region outage takes the whole environment with it, not only the servers. What has to exist in a second region before the day you need it.
- Lesson 2 · 3 min readService Quotas, Machine Images and Other Gaps in the Recovery RegionA recovery region can have the right templates and still fail on quotas, missing machine images or drift. What AWS says to check, and how often.
- Lesson 3 · 3 min readDependency Capture: What a Tool Must Record to Rebuild an ApplicationAn application comes back only if its dependencies come back first and in the right order. What dependency capture means and what each tool documents.
How does the environment come back?
The mechanics of recovery, from the first failover to the return home.
- Lesson 4 · 3 min readFailover or Rebuild: Two Ways Cloud Infrastructure Comes BackSome tools keep a copy running elsewhere and switch to it; others rebuild the environment from a snapshot or code. How the two approaches differ and what each costs.
- Lesson 5 · 3 min readFailback: Getting Back to the Primary Region After a FailoverFailing over is half of a recovery. Failback returns service to the primary region without losing the changes made in the meantime. What to plan and what vendors publish.
- Lesson 6 · 3 min readRunbooks and Automated Recovery: What Each Is Good ForA runbook tells people what to do; automated recovery does it for them. Where each fits in a cloud DR plan, and why you still need a runbook when a tool does the rebuild.
What else has to come back?
Clusters, SaaS settings and other clouds that a plan built around servers can miss.
- Lesson 7 · 3 min readKubernetes Clusters in a Cloud DR PlanA Kubernetes cluster is both cloud infrastructure and a platform with its own configuration. What has to come back, and what the tools here document.
- Lesson 8 · 3 min readSaaS Configuration in a Cloud Recovery PlanIdentity providers, DNS and CDN services, observability and paging tools hold configuration your cloud depends on. Why it belongs in a recovery plan and what is documented.
- Lesson 9 · 3 min readRecovering Across AWS, Azure and Google Cloud: Coverage by ToolCloud DR tools rarely cover every cloud equally. Which of the tools scored here document AWS, Azure and Google Cloud for recovery, and what to ask before you rely on a cloud logo.