What does Operations Tracking show?

Operations Tracking shows every GraphQL operation executed against your federated graph, identified by name and type (Query, Mutation, or Subscription). Unnamed operations are shown as "Unnamed Operation". For each operation, you can see performance metrics, client usage, and deprecated field indicators.

What sort options are available?

You can sort by request count (default), latency, or error rate, in ascending or descending order. Sorting by latency reveals the slowest operations. Sorting by error rate reveals the most problematic ones.

How do I find operations using deprecated fields?

Enable the deprecated fields filter in the operations list. Only operations that use deprecated schema fields will appear. This is the starting point for planning safe schema migrations.

Can I navigate from an operation to its traces?

Yes. Clicking through to traces from an operation pre-applies the operation as a filter in Trace Analytics. You see only the traces for that specific operation.

Can I search for a specific operation?

Yes. You can search by operation name or by operation hash.

Does Operations Tracking require any client setup?

No setup is required to see operations. For client usage data within the detail panel, configure client identification headers (GraphQL-Client-Name and GraphQL-Client-Version) on your client applications.

Operations Tracking for Federated GraphQL | Cosmo by WunderGraph

Q: What performance charts does the detail panel show?

The operation detail panel shows visual charts for request rate, P95 latency, and error percentage trends for the selected operation.

The problem

Managing a GraphQL API without an operation inventory is flying blind

You cannot optimize, migrate, or safely evolve an API you cannot see in full. Teams need a complete, sortable view of what is running.

No complete picture of what is running

Without an operation inventory, teams have no way to know which queries, mutations, and subscriptions are executing against their federated graph. Optimization efforts are guesswork.

Slow or failing operations are hard to find

With nothing to sort or filter by latency or error rate, identifying the worst-performing operations requires manual investigation across logs and metrics. By the time you find the culprit, users have already been affected.

Schema migrations proceed without usage context

Deprecated fields may still be in active use by specific operations and clients. Without seeing which operations use deprecated fields, migrations break things unexpectedly.

Our solution

A complete, sortable operation inventory

Operations Tracking centralizes every GraphQL operation in a two-panel interface: a searchable list on the left and a detail view on the right. Sort, filter, and navigate to traces — all from one place.

What you can do

Operations Tracking provides a complete, searchable inventory of every GraphQL operation executed against your federated graph.
Operations are identified by name (or shown as "Unnamed Operation") and type: Query, Mutation, or Subscription.
Sort by request count to find the busiest operations. Sort by latency to find the slowest. Sort by error rate to find the most problematic.
Filter by deprecated field usage to see which operations will be affected by schema evolution.
Filter by client name to see which operations a specific application executes.
Click any operation to view client usage, deprecated field indicators, and performance charts.
Navigate directly to traces with pre-applied filters for deeper investigation.

From operation list to root cause, without leaving Cosmo Studio.

Operations Tracking

Before & After

Before Cosmo	With Cosmo
Unknown operation inventory	Complete list of all executed operations
Guessing which operations cause issues	Sort by latency or error rate to find problems
Manual tracking of deprecated field usage	Automatic indicators for operations using deprecated fields
Disconnected metrics and traces	Navigate directly from operations to traces with pre-applied filters

Sort and filter options

Find what matters, fast

Sort by request count: Default sort. Find your busiest operations.
Sort by latency: Surface the slowest operations for optimization.
Sort by error rate: Find the most problematic operations during incidents.
Deprecated fields filter: Show only operations that use deprecated schema fields.
Client name filter: Show operations executed by a specific client application.
Search by name or hash: Find a specific operation by name or operation hash.

How Operations Tracking works

01

Count, latency, or error rate.

Sort

Sort by request count (default), latency, or error rate in ascending or descending order. Cross-referencing high-traffic and high-latency operations identifies the best optimization targets.

02

Deprecated fields. Client name.

Filter

Filter by deprecated field usage to find operations that need migration attention. Filter by client name to see what a specific application is executing. Search by operation name or hash.

03

Client usage. Deprecated fields.

Inspect

Click any operation to open the detail panel. See client usage with request counts, a list of deprecated fields used by the operation, and performance charts for request rate, P95 latency, and error percentage.

04

Pre-filtered traces. Operation content.

Debug

Navigate directly from an operation to the Analytics Traces page with pre-applied filters for that operation. Inspect the complete GraphQL operation content with syntax highlighting in a modal.

What is included

Built into every Cosmo plan

Available on Free, Pro, and Enterprise.

Complete operation inventory

Every query, mutation, and subscription executed against your federated graph. Named operations are listed by name; others appear as "Unnamed Operation".

Performance prioritization

Sort by request count, latency, or error rate to direct optimization effort where it has the most impact. Sort direction is configurable.

Deprecated field tracking

Visual indicators show which operations use deprecated schema fields. The deprecated fields filter narrows the list to only those operations — the starting point for any schema migration.

Integrated debugging

Navigate from any operation directly to Trace Analytics with that operation pre-applied as a filter. Inspect the full operation content with syntax highlighting.

See every operation running in your federated graph

Deploy the Cosmo Router and operations start appearing in Cosmo Studio automatically.

Start Free Read the Docs

Find slow, failing, or deprecated operations before they become incidents