Metrics

Track requests per second, latency percentiles, and runtime resource usage for each Unkey deployment. Monitor performance with built-in metric charts.

Unkey Deploy is in public beta. To try it, open the product switcher in the top-left of the dashboard and select Deploy. During beta, deployed resources are free. We’re eager for feedback, so let us know what you think on Discord, X, or email [e-mail].

Each deployment’s overview page includes real-time charts for throughput, latency, CPU, and memory across all instances. Per-instance runtime metrics (CPU, memory, disk, and network) are also available from the deployment’s network view. Navigate to a deployment from your project’s Deployments tab to view them.

Deployment overview charts

Four metric cards at the top of the deployment overview show the last six hours of activity, aggregated across every instance in the deployment. Each card updates automatically every ten seconds.

Requests per second

The RPS chart shows the number of requests your deployment handles over time. Use it to spot traffic spikes, confirm scaling behavior, or correlate with incidents.

Latency

The latency chart displays response time distributions over time. You can switch between p50, p75, p90, p95, and p99 percentiles. Check the Requests tab to find individual slow requests and inspect their latency breakdown.

CPU

The CPU chart shows used CPU as a percentage of the deployment’s total allocated CPU. Hover over a point to see the underlying millicore value. Use this to confirm headroom against your CPU allocation and to predict when autoscaling will kick in.

Memory

The memory chart shows used memory as a percentage of the deployment’s total allocated memory. Hover over a point to see the raw value in MiB. Use this to detect leaks, validate sizing, and decide when to raise the memory limit.

Per-instance runtime metrics

Open the Network tab on a deployment and click an instance node to open the details panel. The panel shows live charts for the selected instance, refreshing every few seconds:

• CPU. Used vs. allocated CPU in millicores.
• Memory. Used vs. allocated memory in bytes.
• Disk. Used vs. allocated disk in bytes. Only shown when the instance has provisioned disk storage.
• Network. Egress and ingress traffic over time.

Use the time window selector at the top of the panel to switch between the past 15 minutes, 1 hour, 6 hours, or 24 hours. When you select a deployment node instead of a single instance, the charts aggregate across all instances in the deployment.