Diagnosing CI performance problems is frustrating when you can't see what's happening on the machine. With the new resource usage feature, you can peer into exactly that: whether one build ballooned to 7GB, whether four tasks at 2GB each collectively pushed the box over the edge, or whether memory had been climbing since minute one.
Resource usage records CPU and memory while your tasks run and charts it back to you in Nx Cloud, down to the individual process.
Now available to every organization
Resource usage needs Nx 22.1 or later. Metrics are collected in CI only. Runs on your local machine don't report resource usage.
Resource usage started as an enterprise-only feature, but now we've expanded it to all orgs as an add-on. Newly created organizations have it enabled by default, so a fresh workspace reports metrics from its first CI run.
Each resource usage report collected costs 10 additional credits. With Nx Agents, every agent in a CI Pipeline Execution will generate one report each. For non-distributed runs, all reports generated on the same machine count as a single 10 credit charge, however many runs happen there. You can turn collection off at the organization or workspace level to skip the cost entirely (see enabling resource usage below). Full details live in the credits pricing reference.
New insights you can derive for your CI
- Debug OOM kills - trace an out-of-memory failure to the exact task responsible.
- Find memory-hungry tasks - spot the one project that eats memory in parallel, then lower its parallelism instead of slowing your entire pipeline down.
- Detect memory leaks - look for tasks where memory climbs steadily rather than spiking and releasing.
- Compare performance of tasks before and after configuration changes - check whether a parallelism tweak or dependency bump quietly doubled your build's memory footprint.
New views to provide more insight into your CI
There are two different locations within the Nx Cloud app where we provide insight into resource usage. Both views share the same controls and let you download the raw resource usage data as CSV.
Runs with Nx Agents
Open any CI pipeline execution and go to the Resource usage tab.
The summary table shows you each agent that ran as part of your CI Pipeline Execution (CIPE) along with their average and maximum CPU and memory. If any agent restarted due to memory overload, you can view the resource usage across agent attempts like shown with Agent 0 in the screenshot above.
Click on any agent to see the resource usage of every task that ran over its lifetime. Separate memory and CPU charts plot utilization, with reference lines marking the machine's capacity and the run's peak resource usage.
Runs without Nx Agents
For a run that isn't distributed, open the run details and go to the Resource usage view.
The memory and CPU charts show the respective resource usage for each task in the Nx command that was executed.
Enabling resource usage
The new resource usage feature is an add-on that is automatically enabled for new organizations and available for any org to toggle on or off. You have the ability to control this feature across your entire org, or more granularly for specific workspaces.
Organization level
You'll find the toggle to control resource usage for your entire organization under Organization Settings > Add-ons. Turning it off here stops collection for every workspace.
Workspace level
Once the add-on is enabled for the org, each workspace decides which of its runs report metrics. Head to your Workspace Settings > General and look for the Resource usage reports card, which has two independent switches:
- Distributed runs - metrics for runs distributed across Nx Agents.
- Non distributed runs - metrics for runs from non-distributed Nx commands.
Flipping a switch off stops new reports from being generated. Anything collected earlier stays viewable for as long as the add-on is active for the organization, so turning collection off doesn't erase your history.
Try it out
The next time CI misbehaves, you won't have to guess. Open the run in Nx Cloud, check the Resource usage view, and find the task that actually caused it.
