New Atlas Metrics UI

Cluster-Level Navigation Changes

With the Metrics page now available at the cluster level, related monitoring pages are now available from the cluster-level secondary navigation to improve discoverability:

Real Time

The Real-Time Performance Panel is now accessible from cluster-level navigation. It includes a node filter to view real-time information per node.

Last Ping and Daily Pings

These pages, previously dependent on metrics page navigation, are now available in the cluster-level secondary navigation.

The consolidated navigation makes it easier to access all monitoring and performance tools from a single location within each cluster.

Chart Legend and Layout

Each chart group displays a legend at the bottom of the group. The legend remains sticky on the page as long as the chart group is in view. Tooltip values appear in descending order for easier comparison.

Filtering and Selection

Filter by Shards and Nodes

Filter metrics to specific shards or individual nodes within your cluster.

For both sharded clusters and replica set clusters:

• The filter defaults to primary series for first-time users.

• Your selection is retained for subsequent visits.

• Shard and node names use abstracted naming (not full hostnames).

• Select individual nodes or view metrics across all nodes.

• The legacy UI also has a node filter with similar capabilities.

For search nodes with decoupled search architecture:

• Search nodes appear separately in the selector.

• Search node names follow the pattern: search-{shardNo}-{randomString}.

• Search nodes are identified with a distinct icon.

• Select search nodes individually or use Select All for quick selection.

Date and Time

Set custom date and time ranges for metrics display. The interface supports both preset time ranges (last hour, last day, and so on) and custom date or time selection.

Select Metrics

Add or remove metric groups from the dashboard using the metrics selector.

• Metrics organization: The metrics selector organizes metric groups into categories, all expanded by default.

• Default selection: The Essential category displays by default for first-time users.

• Persistent selection: The UI retains your selected metrics for subsequent visits.

• Add metric groups: Select a metric group to add a chart card with that group's metrics to the dashboard. Each group typically contains multiple charts.

• Individual chart selection: Use the group-level selector within a chart card to add or remove individual charts from the group.

• Metric descriptions: Hover over any metric to see its description in the right panel. The default state shows category descriptions.

• Search functionality: Search filters all categories except Essential, which contains metrics duplicated from other categories.

Display Data

Control how metric values are aggregated and displayed:

• Sum: Display the sum of metric values.

• Average: Display averaged values. The legend, chart, and tooltip update accordingly.

• Max: Display maximum values.

• Sum by shard: Display the sum of metric values grouped by shard.

• Average by shard: Display averaged values grouped by shard.

• Max by shard: Display maximum values grouped by shard.

The "by shard" options aggregate data at the shard level and display one series per shard based on your host selections. Changing the display data setting updates all chart displays, legends, and tooltips accordingly.

Event Overlays

Toggle event overlays to display database events directly on metric charts. The new Metrics UI includes the same event overlays available in the legacy UI:

• Host Restart: Displays on the time series line for the affected node.

• Elections: Displays on the time series line for the affected node.

Event details appear in tooltips when you hover on the specific point.

Display Controls

Settings Drawer

Access chart granularity settings through the settings drawer.

Hardware Metrics

Hardware metrics display system-level performance data for the hosts supporting your cluster:

• CPU utilization and process CPU.

• Memory usage and available memory.

• Disk IOPS, disk space, and disk latency.

• Network throughput (bytes in or out).

• System CPU and normalized CPU metrics.

MongoDB Metrics

MongoDB process metrics show database server performance:

• Operations per second (opcounters).

• Connections (current and available).

• Query execution metrics.

• Replication lag and oplog metrics.

• Cursor statistics.

• Document metrics for inserted, updated, deleted, and returned documents.

• Cache metrics.

• Queue metrics.

• Assert metrics.

• Page fault metrics.

• MongoDB memory metrics.

• MongoDB network metrics.

Database Statistics (DBStats)

Database-level statistics provide storage and collection metrics:

• Database and collection sizes.

• Index sizes.

• Document counts.

• Data size and storage size.

• Number of collections, indexes, and views.

Database Selector: Use the database selector dropdown to choose which database to view metrics for. The local database is selected by default.

Search Node Metrics

The new UI displays the full suite of search metrics regardless of search deployment type:

• Search node hardware metrics.

• Search-specific process metrics.

• Search index performance data.

• Resource utilization for search workloads.

• Search opcounters.

• Replication lag for search nodes.

• Index statistics (total number and performance metrics).

For dedicated search nodes (decoupled search architecture), search nodes appear as selectable nodes in the topology selector. You can view hardware metrics for these nodes in addition to search-specific metrics.

For colocated search processes, search metrics display based on the selected host. When you select a host with a colocated mongot process, the search metrics reflect that process.

Analytics Node Metrics

Analytics node visibility includes:

• Analytics node hardware performance.

• Query performance on analytics nodes.

• Resource consumption metrics.

Sharding Metrics

For sharded clusters, a dedicated Sharding Metrics category displays:

• Shard data distribution.

• Shard collection counts.

• Cluster collection counts.

• mongos router metrics.

Sharded clusters can aggregate metrics by shard using the "by shard" display options. The same charts are available for both sharded and replica set clusters.

Node and Primary or Secondary Filtering

The new UI provides consistent node selection across both sharded and replica set clusters:

• Select individual nodes or multiple nodes simultaneously.

• Filter by shard membership.

• Visual indicators show node roles: primary, secondary, or arbiter.

• Both topologies use the same filtering interface with additional hierarchy for sharded clusters.

The legacy UI's Toggle Members control is replaced with the Shards and Nodes filter.

Primary or Secondary Series: Select between viewing metrics for primary or secondary nodes across elections. This option allows you to track metrics consistently for a specific node role, even when nodes change roles during elections. For example, selecting "Primary Series" displays metrics for whichever node is currently primary, maintaining continuity across role changes.

Node Naming: In the new Metrics UI for Atlas, shard and node names are abstracted rather than showing full hostnames. Abstracted names improve readability and focus on logical cluster structure. For example:

• Shard names appear as <cluster-alias>-shard-0, <cluster-alias>-shard-1, and so on, where the cluster alias may be abbreviated if long.

• Search nodes appear as search-{shardNo}-{randomString}.

• Node names use shortened identifiers instead of full Atlas hostnames.

Chart legends and tooltips use these abbreviated node names for cleaner displays. You can copy full hostnames using the copy button in the chart legend.

Event Overlays

Event overlays can be toggled on or off and synchronize across all charts. The new Metrics UI includes Host Restart and Elections events that display on the time series lines for affected nodes.

Chart Interaction

Enhanced chart interactions include:

Zoom Controls:

• Zoom in: Click and drag across a time range on any chart to zoom in.

• Zoom out: Use the top-level date and time or time range selectors to adjust the view.

Chart Group Actions:

• Expand/Collapse: Expand a chart group to show all metrics, or collapse it to save space.

• Metric Description: Click the info icon to open a modal with detailed metric descriptions.

• Close Chart Group: Remove an entire group of related metrics from the dashboard.

• Collapsed State: When collapsed, chart groups show summary information while taking up minimal space.

Individual Chart Actions:

• Close Chart: Remove an individual chart from a chart group.

Performance Improvements:

• Better mobile and tablet responsiveness.

• Smoother chart rendering and transitions.

• Optimized for scrolling through many charts.

Empty State Messaging

When database statistics collection is disabled, the new UI displays:

• Clear empty state messages explaining why data is unavailable.

• Direct links to enable collection in project settings.

• Guidance on the performance impact of enabling collection.

Supported Deployment Types

• The new Metrics UI is available for standard Atlas dedicated clusters (M10 and above).

• Free tier (M0) and Shared tier (M2 or M5) clusters use separate metrics interfaces.

• Flex clusters have their own dedicated metrics view.

Environment Scope

The new Metrics UI applies only to:

• Atlas cloud-hosted deployments.

• Project-level metrics pages.

It does not affect:

• Cloud Manager metrics interfaces.

• MongoDB production government environments.

• Atlas Internal metrics.