maxInstances

MaxInstances is a configuration parameter used in software systems to cap the number of simultaneously active instances of a given component, such as a service, function, or worker. The setting is commonly found in serverless platforms, containerized deployments, and microservice architectures, where controlling concurrency helps manage resource usage and predict performance.

When a request arrives and the active instance count equals MaxInstances, new requests are either queued, rejected with a throttling error, or redirected to existing instances depending on the platform. The exact behavior depends on the runtime and configuration; some systems prefer backpressure by delaying execution, others opt for immediate failure to preserve latency guarantees.

Configuration and naming vary; the parameter may be called MaxInstances, maxInstances, max_concurrency, or similar, and can be set as a static value or dynamically adjusted by autoscaling policies. In some environments, MaxInstances interacts with autoscaling thresholds, warming periods, and startup latency.

Trade-offs: a higher limit allows greater throughput but increases resource consumption and the potential for contention; a lower limit reduces resource usage but can raise latency or cause request failures under load. Effective use requires workload profiling, monitoring of latency and error rates, and alignment with autoscaling rules.

In practice, MaxInstances is used in cloud functions, API gateways, and batch processing pipelines to enforce capacity planning and predictable performance. See also concurrency limits, throttling, and autoscaling.