Is the AWS Bedrock Academy free?

Yes. The full course is free and includes 25 modules, 52 lessons, 16 labs, and quiz-based progression with no sign-up required.

How is Infinity AWS Bedrock Academy different from AWS Skill Builder?

Unlike AWS Skill Builder which often gates advanced hands-on labs behind paid subscriptions or AWS Builder ID logins, Infinity Tech Stack's AWS Bedrock Academy is 100% free, ungated, and browser-first with zero sign-up required. It includes 25 modules, 52 lessons, 16 hands-on labs, and 125+ quiz checks covering the Converse API, Bedrock Guardrails, AgentCore, and multi-model evaluations without requiring active AWS account credentials.

Who is this course for?

Developers, architects, and platform teams building AI systems on AWS Bedrock from MVP to production scale.

What does this Bedrock course cover?

The academy covers runtime inference, prompt engineering, guardrails, knowledge bases, agents, evaluation, observability, security, compliance, resilience, incident response, and production launch discipline.

Failover runbooks should be:

Regularly tested and versioned — Tested runbooks are essential for reliable incident response.

Regional Failover Patterns Tutorial | Multi-Region Resilience

← Back to Dashboard

1. Regional Failover Patterns2. Chaos Testing for AI Services

Regional Failover Patterns

📚 Multi-Region Resilience⏱ 10 min⭐ 115 XP

Failover Is a Product Requirement

Use active-passive or active-active strategies depending on your latency and compliance profile. Bedrock adds one twist classic DR plans miss: the same model may not exist - or behave identically - in your failover region, so model equivalence is part of your failover design.

Strategy Options

Strategy	Shape	Cost/complexity	Fit
Cross-region inference profiles	AWS routes across a geography automatically (us./eu. prefixed IDs)	Lowest - built in	First line of defence for throughput + regional blips
Active-passive	Standby region with config + quotas pre-provisioned; DNS/flag flip to fail over	Medium	Most teams; clear runbook, bounded spend
Active-active	Both regions serve traffic continuously	Highest - double quotas, sync complexity	Hard availability targets, global user base

The Model Equivalence Map

# failover-map.yaml
primary:   us-east-1
secondary: us-west-2
models:
  amazon.nova-pro-v1:0:      same-id          # verified available
  anthropic.claude-sonnet-*: us.profile       # cross-region profile covers both
  custom.triage-ft-v2:       nova-pro + prompt-v3-fallback   # custom models don't failover - plan the substitute!

Custom models and Provisioned Throughput are region-bound - your failover plan must name the substitute and accept its eval-scored quality delta in advance.

What Must Be True Before You Can Fail Over

Model access granted in the secondary region (it is account+region scoped - module 2).
Quotas raised in the secondary - default TPM in an unused region will not absorb your production load.
Guardrails, prompts, KB replicas deployed - inference without your safety config is not a failover, it's an incident.
Data residency audited - an EU workload failing over to us-east-1 may trade an outage for a compliance breach.
Runbooks tested and current - untested failover plans fail at the worst possible moment (next lesson: chaos drills).

Quota trap: teams provision failover infra but forget service quotas. The failover "works" - at 5% of required throughput. Request secondary-region quota increases as part of DR setup, not during the incident.

🧪 Knowledge Check

Press 1-4 to select1 of 2

Why maintain region model equivalence maps?

For aesthetics

To ensure predictable behavior during failover

To remove IAM

To avoid backups