3 min read
Azure Chaos Studio Workspaces can test a complete failure story against a controlled scope. The result is evidence of how the application recovered, not another statement of intent.
Owner: Cloud operations
A resilient architecture can still fail at the hand-off between compute, DNS, identity, data and application logic. The only reliable way to find that gap before a customer does is to test a controlled failure and measure the recovery.
What changed
Microsoft has placed Azure Chaos Studio Workspaces in public preview. A Workspace is scoped to a subscription, resource group or service group. It discovers the relevant resources and recommends named failure scenarios based on what is in scope.
The scenarios reflect outage patterns rather than isolated technical faults. Examples include Availability Zone Down, DNS Outage, database failover, Microsoft Entra ID Outage, Cache Stampede and messaging disruption. A scenario can exercise more than one layer, which matters because an application may survive the platform event but fail to reconnect, retry, preserve data or route traffic correctly.
After a run, Microsoft says the Workspace produces a structured drill report. It records what was injected, which resources were affected, how the recovery timeline unfolded, which monitoring signals changed and where behaviour differed from the expectation.
Why this matters to a South African organisation
Business continuity plans often describe what should happen. Customer confidence depends on what does happen. A tested recovery path gives technology leaders evidence for service reviews, change records, audit packs and executive risk discussions.
This does not mean experimenting carelessly in production. Chaos Studio injects real disruption. Public preview status increases the need for defined scope, permissions, change control, abort conditions and rollback discipline.
What to do this week
Choose one customer journey.
Start with a path the business understands, such as sign-in, order capture, payment processing or access to a core application.
Use a safe environment.
Begin in preproduction. Confirm that the data, dependencies and monitoring are representative enough to produce a useful result.
Write the hypothesis.
State what should continue working, what may degrade, how quickly recovery should occur and what data must remain correct.
Define the blast radius.
Scope the Workspace narrowly. Approve the affected subscription or resource group, identities, permissions and stop conditions.
Run one scenario.
Zone Down is a strong first drill because it exposes routing, database, DNS and application-retry assumptions.
Capture the evidence.
Compare the drill report and Azure Monitor signals with the expected recovery objective. Record every difference as an owned action.
Evidence to retain
Keep the approved hypothesis, change record, scoped resources, permissions, abort conditions, monitoring baseline, drill report, recovery timeline and corrective actions. That is stronger evidence than a diagram labelled highly available.
The Braintree view
Resilience is an operating discipline. Redundancy, failover and retry settings are inputs. Recovery evidence is the outcome. Prove one customer path safely, fix what the drill exposes and repeat until the result matches the promise.
CONNECTED EXPERTISE
