Monitoring and performance metrics
PAI-Instant Service provides multi-level Grafana dashboards (instant service, template, and instance) and benchmark performance data from load testing. This topic describes how to access the dashboards and the capacity and throughput you can expect in high-concurrency workloads such as RL training and agent evaluation.
Grafana dashboards
On the instant service workbench, template details page, or instance details page, click Open monitoring page to open the Grafana dashboard for that scope.
The dashboards on the instant service workbench and the template details page display the following metrics:
Instant Service:
Instance Count: total instance count, broken down by running and paused.
Instance Creation Latency: instance creation latency (max, avg, P50, P90, and more).
Top Instance Templates: the most-used templates by usage count.
Requested CPU Cores / Memory / Storage: total resource requests across all instances.
Gateway API:
Instance Creation/Deletion/Connection QPM: queries per minute (QPM) for the instance creation, deletion, and connection control-plane APIs.
Instance Operation Count: cumulative count of instance creation, deletion, and connection operations over the selected time range.
Proxy Request: QPS, latency, and error rate for the Gateway proxy API that routes requests to instance services.
The instance details page dashboard shows runtime metrics for a single instance, including CPU and memory utilization, network, disk, and process status.
Performance metrics
PAI-Instant Service has been load-tested at large-scale concurrency. The following benchmark data helps you estimate capacity and throughput for high-concurrency workloads such as RL training and agent evaluation.
Test environment
Cluster: PAI Instant Service compute cluster.
Node pool size: 300 nodes.
ecs.c6.13xlarge(52 vCPU / 91 GiB) × 277ecs.c6.26xlarge(104 vCPU / 183 GiB) × 23
Test date: April 2026.
Test methodology
A client script submits instance creation requests to the control plane in bulk. The test starts when the script begins sending requests and ends when the last instance starts. The elapsed time is the concurrent instance startup duration.
Template spec
Code Interpreter (1 CPU / 2 GiB memory / 8 GiB disk)
Test results
At 10,000+ concurrent instances, PAI-Instant Service sustains an overall QPS above 750 with a per-node P99 startup latency under 600 ms, meeting the demands of large-scale agent evaluation and RL rollout workloads.
Instances created | Concurrent startup duration | Overall QPS | Avg node latency | Node P50 latency | Node P99 latency |
10,000 | 13.23 s | 755.74 | 396.02 ms | 395.81 ms | 502.73 ms |
20,000 | 26.17 s | 764.23 | 457.62 ms | 437.96 ms | 582.07 ms |
Note: The concurrent startup duration is the total time for all instances to start, which includes client-side concurrency limits and other constraints. For single-instance startup time, refer to the node latency figures.