AutoscalingSafetyLimits Interface

Safety limits for the autoscaler. If any limit is exceeded, the autoscaler reverts the increase and stops proposing new increases.

interface AutoscalingSafetyLimits {
    maxCpuRatio: number;
    maxRssRatio: number;
    maxErrorRate: number;
    maxP95Ms: number;
    maxP99Ms: number;
    maxQueueFull: number;
    maxQueueTimeout: number;
}

Properties

maxCpuRatio: number

Maximum CPU usage (0 to 1). Default: 0.85.

maxRssRatio: number

Maximum RSS memory usage (0 to 1). Default: 0.8.

maxErrorRate: number

Maximum error fraction (0 to 1). Default: 0.001.

maxP95Ms: number

Maximum p95 latency in ms. Default: 500.

maxP99Ms: number

Maximum p99 latency in ms. Default: 1000.

maxQueueFull: number

Maximum number of rejections per full queue. Default: 0.

maxQueueTimeout: number

Maximum number of rejections due to a queue timeout. Default: 0.