Proactive Work: Launches, Resilience, and Cost
Module 3 · Operational Excellence

Proactive Work: Launches, Resilience, and Cost

How TAMs prevent problems: operational reviews and cadence calls, launch readiness with Countdown, lifecycle and quota management, game days, and cost-optimization reviews with Savings Plans, right-sizing, and storage tiering.

Proactive Work Is the Job

The best incident is the one that never happens. Proactive work is where a TAM earns the customer’s trust, and it’s what interviewers want to hear about in your stories.


The Operating Cadence

Most TAMs run a regular rhythm with each customer:

CadencePurposeTypical content
Weekly or biweekly callKeep things movingOpen cases, upcoming events, AWS Health notices, action items
Monthly operational reviewSpot trendsCase trends, incident review, Trusted Advisor risks, cost trend
Quarterly business review (QBR)Align with leadershipOutcomes delivered, risks, plans for next quarter (Module 4)

The cadence should produce action items with owners. A meeting that ends with “great sync, talk next week” didn’t do its job.


Launch and Event Readiness

When a customer plans a big event (a product launch, a migration cutover, a seasonal peak), the TAM leads readiness, often through AWS Countdown:

  1. 1 Define

    Date, expected peak, success criteria

  2. 2 Review

    Architecture, single points of failure

  3. 3 Quotas

    Limits, capacity, scaling plans

  4. 4 Test

    Load test with real traffic shape

  5. 5 Prepare

    Runbooks, alarms, contacts

  6. 6 Run + review

    Standby support; lessons learned

Questions a TAM asks before any launch:

  • What’s the expected peak, and what happens at 2× that?
  • Have you checked service quotas for every service in the request path?
  • Do Auto Scaling policies react fast enough, or do you need to pre-scale?
  • What’s the rollback plan? Who decides to use it?
  • Who’s on call, and does AWS have the escalation contacts?

Lifecycle and Risk Management

Many outages come from things that were announced months in advance:

  • End of standard support for database engine versions (RDS, Aurora, ElastiCache), EKS Kubernetes versions, and Lambda runtimes
  • Certificate expirations
  • Instance retirements and scheduled maintenance

A TAM watches AWS Health and keeps a lifecycle tracker for each customer: what’s coming, which resources are affected, who owns the upgrade, and the deadline. It’s unglamorous work, and it prevents a lot of outages.

Game days

A game day is a planned exercise that simulates a failure in a controlled way: kill an AZ, fail over a database, throttle a dependency. Teams then check whether alarms fire, runbooks work, and people know their roles. AWS Fault Injection Service supports these experiments. TAMs often help customers plan their first game day.


Cost Optimization

Cost reviews are one of the most visible ways a TAM delivers value. The main levers:

LeverWhat it isTypical finding
Right-sizingMatch instance size to actual useInstances averaging 8% CPU (check Compute Optimizer)
CommitmentsSavings Plans or Reserved Instances for steady usageA steady baseline paid entirely at On-Demand rates
SpotSpare capacity at deep discount for interruptible workBatch or CI jobs on On-Demand
Storage tieringS3 lifecycle rules, Intelligent-Tiering, gp2 → gp3Years of logs in S3 Standard
WasteIdle resourcesUnattached EBS volumes, idle load balancers, old snapshots
ArchitectureChange how it’s builtData transfer costs from cross-AZ chatter, or NAT Gateway processing
Modern hardwareGraviton instancesBetter price-performance for compatible workloads

Order matters: right-size and remove waste before buying commitments. Otherwise the customer commits to paying for capacity they didn’t need.

Illustrative monthly bill: shrink usage first, then commit to what’s left100−10−12−1860Starting billRemove wasteRight-sizeCommitOptimized billidle, unattachedCompute OptimizerSavings Plans, RIs
⚠️ Common Pitfall: Optimizing Cost at the Expense of Reliability

Common mistake: a candidate’s story is “I cut the bill 40% by moving everything to Spot and removing the standby database.”

Cost is one pillar of six. A good TAM presents trade-offs: “This saves $200K a year but raises your RTO from 2 minutes to 30. Is that acceptable for this workload?” The customer decides, with full information.


Key Takeaway

Proactive TAM work runs on a cadence that produces owned actions. It includes launch readiness (Countdown, quotas, load tests, rollback plans), lifecycle tracking through AWS Health, game days to test recovery, and cost reviews: right-size and remove waste first, then commit, and always state the reliability trade-off.

🧪 Try It: Build a Cost Review Agenda

A customer spends $400K a month: 60% EC2 (all On-Demand), 20% RDS, 10% S3, 10% data transfer and NAT Gateway. Usage has been steady for a year.

Draft a 30-minute cost review agenda. For each item, say what data you’d bring, the likely savings lever, and any reliability trade-off. Order the items so the customer doesn’t commit to waste.

🧠 Knowledge Check

1. Why should right-sizing come before buying Savings Plans?

2. Where would a TAM track an upcoming database engine end-of-support date?

3. What is the main purpose of a game day?

4. Before a launch, which question helps prevent the most common launch-day surprise?