How does auto-scaling work and what metrics should be used?
нажми, чтобы перевернуть
midscalability
Automatically adjusting the number of instances based on load metrics.
Auto-scaling adds/removes instances based on rules: CPU > 70% — scale out, CPU < 30% — scale in. Metrics: CPU, Memory, Request Count, Queue Depth, Custom metrics. Cooldown period prevents flapping. Predictive scaling based on historical patterns.
