Join us at MongoDB.local London on 7 May to unlock new possibilities for your data. Use WEB50 to save 50%.
Register now >
Docs Menu
Docs Home
/ /

Adaptive Operation Rate Limiting

The Adaptive Rate Limiter for Operations is an Intelligent Workload Management (IWM) policy in Atlas. IWM is a dynamic resource manager that provides real-time workload monitoring and automated safeguards to maintain high availability under load. The Adaptive Rate Limiter for Operations dynamically adjusts the rate at which MongoDB accepts and rejects operations that your applications send to a mongod when a cluster is overloaded.

MongoDB considers a node overloaded when the number of incoming operations is large enough to cause total or near-total outage. MongoDB computes overload from metrics like CPU utilization, queue depth, operations per second, and latency.

Important

This policy is a load-shedding policy. If this policy is active on your Atlas cluster and your cluster is overloaded, you might see its associated overload errors.

When traffic suddenly spikes, accepting more operations than your node's maximum load can overwhelm your cluster, causing degraded performance, timeouts, and potential failovers. The cluster can take significant time to recover.

The Adaptive Rate Limiter for Operations policy prevents overload by:

  • Limiting the admission rate to what the system can safely handle

  • Maintaining cluster stability and avoiding outages

  • Keeping a portion of operations succeeding with predictable latency

  • Enabling faster recovery from traffic spikes

  • Your Atlas cluster must be running MongoDB 8.3 or later to use this policy. On MongoDB 8.3, this policy is disabled by default. To enable or disable IWM policies, see the IWM settings.

  • This policy is available only for M10+ Atlas replica set clusters.

  • This policy is not available on sharded clusters.

When Atlas runs the Adaptive Rate Limiter for Operations policy on your cluster, it performs the following actions:

  1. Monitors for overload

  2. Determines a safe admission rate

    • When the system approaches overload, Atlas computes a maximum safe rate at which it can admit new operations on each node, based on recent conditions.

  3. Admits or rejects operations at the entry point

    • MongoDB admits and runs as usual any operations that arrive within the safe rate.

    • MongoDB immediately rejects any operations that arrive above the safe rate for mongod on each node. MongoDB doesn't queue these operations until they time out.

  4. Adapts the admission rate over time

    • As load decreases and the cluster recovers, Atlas relaxes the rate limit so that it can admit more operations again.

    • When the policy is no longer active, the following informational event appears in the cluster's activity feed:

      "Atlas is no longer regulating the admission rate of new operations."

      To learn more, see the IWM activity feed events.

When the policy is active, some operations in your application fail quickly with an error containing the SystemOverloadedError label. Other operations continue to succeed. This prevents a situation where all operation time out, causing a node crash. To learn more about how to catch overload errors and avoid retry storms, see Overload Errors.

You can use the following methods to track how the Adaptive Rate Limiter for Operations is affecting your workload:

Back

Reliability, Availability, and Workload Management

On this page