🔒 Paid
Autoscaling Failure
Inbound traffic has spiked sharply and continues to climb, but your service isn't adding capacity fast enough to keep up. CPU utilization is pinned near its ceiling, the request queue is growing, and errors are starting to appear. The autoscaler may be reacting too slowly, or it may simply be unable to scale any further — diagnose which it is before the backend falls over.
- Difficulty
- Senior
- Expected Time
- ~14 min
- Category
- Infrastructure & Capacity
Skills
- Diagnosing autoscaling capacity constraints
- Distinguishing reaction-speed tuning from hard capacity limits
- Capacity vs. cost tradeoffs
- Recognizing plausible-but-ineffective mitigations
System Architecture
- Load Balancer
- Autoscaler Controller
- Autoscaling Group
- Backend Service
- Database
This is a Pro scenario. Upgrade to Pro to unlock it.
Upgrade to ProBack to Incidents