Atlas Performance Advisor Auto-Index Evaluation

Parent: Atlas Triage Workflows · Published reference · snapshot 2026-10-15

Depth-first rabbithole dossier for Atlas Performance Advisor Auto-Index Evaluation; source-anchored research pack.

These notes link each claim to its source. Some citations identify archived research reports that are no longer hosted on this site. A published reference means the content is available; it does not certify independent review or accuracy.Read the editorial policy and follow the sources before relying on a claim.

Definitions

[low] IWM is the overload protection system applied to all Dedicated clusters running MongoDB 9.0 or later, but it is not available on Free or Flex clusters. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]

How it works

[low] The Atlas Performance Advisor considers workload activity across all nodes in the replica set, including secondaries, when generating index recommendations. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] When Load Shedding is enabled, Atlas bounds the size of the ingress queue on each `mongod` process; when disabled, it does not bound the queue or reject operations. Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]
[low] If reactive auto-scaling is enabled, Atlas uses the rate of queued operations or operations rejected by Load Shedding to determine when to scale the cluster up. Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]
[low] On sharded clusters, Atlas blocks only writes routed to the affected shard when disk space is critically low. Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]
[low] When used on Flex clusters, the `setServerlessAutoIndexing` and `getServerlessAutoIndexing` commands take the value set by the user but do not toggle Auto-Indexing because it is not available for Flex clusters. Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]
[low] When used on Flex clusters, `upgradeSharedCluster`, `createServerlessInstance`, and `updateServerlessInstance` take the value set by the user but do not toggle Continuous Backup because it is not available for Flex clusters. Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]
[low] When configured, Atlas exports from `mongod`, `mongos`, and audit logs to an AWS S3 bucket every minute. Sources: https://www.mongodb.com/docs/atlas/architecture/current/monitoring-alerts.md [source]

Problems, failure modes and limitations

[low] The Atlas Performance Advisor automatically analyzes logs for slow-running queries and recommends indexes to create and drop, providing suggestions for individual collections ranked by a calculated impact score tailored to the workload. Sources: https://www.mongodb.com/docs/atlas/architecture/current/monitoring-alerts.md [source]
[low] The Atlas Performance Advisor cannot suggest indexes for MongoDB databases configured to use the `ctime` timestamp format. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] By default, Atlas dynamically adjusts the slow query threshold based on execution time, but users can opt out to use a fixed 100-millisecond threshold. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] Atlas clusters with MongoDB Search enabled do not support the Atlas-managed slow query operation threshold. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] For Free clusters (formerly M0), Atlas disables the Atlas-managed slow query operation threshold by default and it cannot be enabled. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] Project Owner access is required to enable or disable the Atlas-managed slow operation threshold. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] The Atlas CLI can return up to 20 namespaces in `<database>.<collection>` format for collections experiencing slow queries. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] The Query Profiler identifies slow queries and bottlenecks, suggests index refinement and query restructuring, and provides visibility into the slowest operations over up to seven days in the Atlas UI. Sources: https://www.mongodb.com/docs/atlas/architecture/current/monitoring-alerts.md [source]
[low] Users can enable the Atlas-managed slow operation threshold for their project using a specific command. Sources: https://www.mongodb.com/docs/atlas/architecture/current/monitoring-alerts.md [source]
[low] Atlas does not recommend setting the fixed slow query threshold lower than 100 milliseconds. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] A specific tool is deprecated and returns an error because Serverless instances are no longer supported as of May 2025. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]

Comparisons and alternatives

[low] When memory runs low, the cluster slows down in a controlled way instead of running out of memory or repeatedly clearing necessary data. Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]
[low] Atlas CloudFormation and CDK support Flex clusters using the `MongoDB::Atlas::Cluster` or `CfnCluster` resources instead of Flex-specific resources, with general support for M2/M5 clusters ending on January 22, 2026. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]

Facts and statements

[low] The Atlas Performance Advisor is only available on M10+ clusters. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] The Atlas Performance Advisor fetches index recommendations from the past 24 hours by default, but users can click Explore Recommendations to look up recommendations for up to the past 5 days if no immediate recommendations are available. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[medium] The Atlas Performance Advisor ranks indexes by Impact, which is determined by the total wasted bytes read by associated operations. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md, https://www.mongodb.com/docs/atlas/performance-advisor/index-ranking.md [source]
[low] The Atlas Performance Advisor does not provide performance suggestions for time series collections. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] Users can only create one index at a time through the Atlas Performance Advisor. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] Atlas always creates indexes for entire clusters; if an index is created while viewing the Performance Advisor for a single shard in a sharded cluster, Atlas creates that index for the entire sharded cluster. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] The Atlas Performance Advisor monitors queries that take longer than 100 milliseconds to execute and groups these queries into common query shapes. Sources: https://www.mongodb.com/docs/atlas/performance-advisor/index-ranking.md [source]
[low] The Atlas Performance Advisor does not suggest indexes that have more than 16 fields. Sources: https://www.mongodb.com/docs/atlas/performance-advisor/index-ranking.md [source]
[low] The Atlas Performance Advisor suggests an index only if the difference between scanned documents and returned documents is greater than 500 for impacted queries. Sources: https://www.mongodb.com/docs/atlas/performance-advisor/index-ranking.md [source]
[low] The Atlas Performance Advisor de-duplicates overlapping indexes before making suggestions, such as suggesting only `{ a : 1, b : 1 }` if `{ a : 1 }` is a prefix. Sources: https://www.mongodb.com/docs/atlas/performance-advisor/index-ranking.md [source]
[low] By default, the Atlas Performance Advisor shows index recommendations for all collections in the cluster, listing them in order of performance impact from greatest to least. Sources: https://www.mongodb.com/docs/atlas/performance-advisor/index-ranking.md [source]
[low] The Real-Time Performance Panel (RTPP) in the Atlas UI provides insights into network traffic, database operations, and hardware statistics at one-second granularity, but is not available via the Atlas Administration API for monitoring, though it can be enabled or disabled via the API. Sources: https://www.mongodb.com/docs/atlas/architecture/current/monitoring-alerts.md [source]
[low] The Atlas UI does not support building indexes with a rolling build for Free clusters and Flex clusters. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] As of February 2025, users can create Flex clusters but can no longer create M2, M5, or Serverless instances via Atlas UI, CLI, API, Kubernetes Operator, Terraform, or CloudFormation. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]
[low] As of January 22, 2026, Atlas no longer supports M2 and M5 clusters and Serverless instances, having migrated all existing M2 and M5 clusters to Flex clusters. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]
[low] Atlas migrated Serverless instances to Free clusters, Flex clusters, or Dedicated clusters according to usage. Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]
[low] Load Shedding is available only on M30+ replica set clusters running MongoDB 9.0 or later. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]
[low] By default, Atlas sets the Load Shedding setting to Atlas Managed, which controls the effective state; on MongoDB 9.0, that state is Disabled. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]
[low] Write blocking and memory tuning are active on M10+ clusters running MongoDB 8.0 or later. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]
[low] Cache eviction is active on M10+ clusters running MongoDB 8.3.3 or later, or MongoDB 8.0.28 or later. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]
[low] Connection rate limiting is active on M10 and M20 clusters running specific MongoDB versions. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]
[low] On Atlas Core clusters, if the primary node runs critically low on free disk space, Atlas blocks write operations (including deletes) on the replica set while continuing to serve reads. Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]
[low] During connection storms, Atlas admits new client connections at a safe rate and rejects excess requests while preserving existing connections. Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]
[low] During overload, Atlas deprioritizes background tasks and long-running operations to help short-running reads and writes continue making progress. Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]
[low] Cluster API endpoints previously compatible with M2/M5 clusters managed Flex clusters as of January 2025 and reached End of Life on January 22, 2026. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]
[low] API endpoints for Serverless Instances managed Flex clusters as of January 2025, and Serverless instances are no longer supported as of March 2025, with these endpoints now only supporting Flex clusters. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]
[low] The HashiCorp Terraform Provider supports Flex clusters, with general support for M2/M5 clusters and Serverless instances ending on January 22, 2026. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]
[low] The Atlas Go SDK supports Flex clusters, with general support for M2/M5 clusters and Serverless instances ending on January 22, 2026. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]
[low] Continuous cloud backup is no longer supported on Serverless instances as of January 2025. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]
[low] General Atlas CLI support for M2/M5 clusters and Serverless instances ended on January 22, 2026. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]
[low] Managing Flex clusters requires Atlas Kubernetes Operator 2.12.0 or later, which is free of serverless dependencies, using `spec.flexSpec` in the `AtlasDeployment` Custom Resource. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/flex-migration.md [source]
[low] Atlas provides alerts for over 200 event types and retains the last 30 days of log messages and system event audit messages. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/architecture/current/monitoring-alerts.md [source]
[low] Atlas configures some alerts by default and provides built-in integrations with third-party tools like Datadog and Prometheus, as well as support for custom metrics tools via the Atlas Administration API. Sources: https://www.mongodb.com/docs/atlas/architecture/current/monitoring-alerts.md [source]
[low] In dev and test environments, it is recommended to configure alerts for clusters inactive for seven or more days to identify candidates for shutdown to save costs. Sources: https://www.mongodb.com/docs/atlas/architecture/current/monitoring-alerts.md [source]
[low] If a cluster experiences an activity spike generating an extremely large quantity of log messages, Atlas may stop collecting and storing new logs for a period of time. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] Atlas automatically cancels rolling index builds that do not succeed on all nodes, removing the index from any nodes where it was successfully built if the build fails on others. Sources: https://www.mongodb.com/docs/atlas/performance-advisor.md [source]
[low] For clusters with FCV 9.0 or later, logs return a `MongoServerError` indicating replica set writes are blocked due to insufficient disk space, while earlier FCV versions indicate user writes are blocked due to disk use threshold exceeded. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/intelligent-workload-management.md [source]
[low] The `Disk read IOPS on Data Partition is` alert condition is raised if the average number of disk read operations per second exceeds the threshold. Sources: https://www.mongodb.com/docs/atlas/reference/alert-resolutions/disk-io-utilization.md [source]
[low] The `Disk write IOPS on Data Partition is` alert condition is raised if the average number of disk write operations per second exceeds the threshold. Sources: https://www.mongodb.com/docs/atlas/reference/alert-resolutions/disk-io-utilization.md [source]

Related concepts

Atlas — is a part of Atlas Performance Advisor Auto-Index Evaluation
Performance — is a part of Atlas Performance Advisor Auto-Index Evaluation
Advisor — is a part of Atlas Performance Advisor Auto-Index Evaluation

Children

No children recorded.

← the whole tree · 3D view· how to read this page