ACK observability uses three Alibaba Cloud services: Managed Service for Prometheus (metrics), Application Real-Time Monitoring Service (ARMS) (traces and application performance), and Simple Log Service (SLS) (logs). Each service bills independently based on usage.
The following table shows which data types each service handles and whether charges apply by default.
| Data type | Service | Charged by default? |
|---|---|---|
| Basic cluster metrics (node, pod, Kubernetes object status) | Managed Service for Prometheus | Depends on edition — see Billing for Managed Service for Prometheus |
| Custom metrics (plug-in-reported metrics) | Managed Service for Prometheus | Yes, based on data volume and storage |
| Application performance data (Java, Python, Go) | ARMS | Yes, when ARMS is enabled for the application |
| Audit logs, control plane logs, container logs | Simple Log Service | Yes, calculated based on usage |
Billing for Managed Service for Prometheus
To collect monitoring data with Managed Service for Prometheus, enable Cluster Monitoring when creating a cluster. Two editions are available: ACK Cluster Monitoring Basic Edition and ACK Cluster Monitoring Pro Edition.
Currently, only ACK Pro clusters support ACK Cluster Monitoring Pro Edition. Other cluster types use ACK Cluster Monitoring Basic Edition by default. ACK is gradually adding Pro Edition support for other cluster types.
Basic Edition vs. Pro Edition billing
| Billing item | ACK Cluster Monitoring Basic Edition | ACK Cluster Monitoring Pro Edition |
|---|---|---|
| Basic metrics | Free by default when a cluster is created. Modifying the default settings may generate additional fees. | Billed based on the number of nodes by default. Modifying the default settings may generate additional fees. See Container Monitoring billing. |
| Custom metrics | Charged based on reported data volume and storage usage. See Prometheus instance billing. | — |
| Cluster resource consumption | Non-managed agents. Each agent occupies 3 CPU cores and 4 GB of memory by default. You manage the agents, and their resource usage generates additional costs. | Managed agents. No additional resource fees. |
ACK Cluster Monitoring Pro Edition also includes built-in Grafana monitoring dashboards, default alert rules for ACK components, and Remote Write and data shipping capabilities via EventBridge. For a full feature comparison, see Differences between the Basic Edition and Pro Edition.
Basic metrics vs. custom metrics
All metrics except basic metrics are classified as custom metrics and billed based on reported data volume and storage usage. To check whether a specific metric is a basic metric, see Container cluster basic metrics. Metrics not listed there are custom metrics. The list of basic metrics is updated continuously as new metrics are added.
The following plug-ins report custom metrics: ack-net-exporter, ack-sysom-monitor, migrate-controller, Knative, ack-arena, and ack-fluid.
Billing for ARMS application monitoring
Fees are generated when Application Real-Time Monitoring Service (ARMS) is enabled for applications running in ACK clusters, such as Java Application Monitoring, Python Application Monitoring, and Go Application Monitoring. Billing is based on ARMS Application Monitoring billing items. For details, see Billing.
Billing for Simple Log Service
Fees are generated when Simple Log Service (SLS) collects and processes logs from ACK clusters — including audit logs, control plane logs, and container logs (pod logs). Charges are calculated based on actual usage and billed directly by SLS. For pricing details, see Billing overview.
Observability features and components
The following table maps each observability feature to the Alibaba Cloud service it uses and the component that reports the data.
To confirm whether a metric collected by Managed Service for Prometheus is a basic metric, see Container cluster basic metrics. Metrics not listed there are custom metrics.
| Module | Feature | Alibaba Cloud service | Data reporting component |
|---|---|---|---|
| Control plane log monitoring | Collect control plane component logs of ACK managed clusters | Simple Log Service | ACK managed clusters |
| CoreDNS log monitoring | Analyze and monitor CoreDNS logs | Simple Log Service | CoreDNS |
| API Server audit log monitoring | Work with cluster API Server auditing and Collect control plane component logs of ACK managed clusters | Simple Log Service | API Server |
| Container auditing | Use Container Behavior Auditing and Collect control plane component logs of ACK managed clusters | Simple Log Service | ack-advanced-audit |
| Container log monitoring | Collect container logs from ACK clusters and Collect control plane component logs of ACK managed clusters | Simple Log Service | logtail-ds component |
| Application monitoring | Java Application Monitoring | Application Real-Time Monitoring Service | ack-onepilot |
| Application monitoring | Python Application Monitoring | Application Real-Time Monitoring Service | ack-onepilot |
| Application monitoring | Go Application Monitoring | Application Real-Time Monitoring Service | ack-onepilot |
| Control plane component monitoring | View control plane component dashboards | Managed Service for Prometheus | API Server, kube-controller-manager, cloud-controller-manager, kube-scheduler, etcd |
| Network monitoring | Use Managed Service for Prometheus to improve the observability of Ingress | Managed Service for Prometheus | Nginx Ingress Controller |
| Network monitoring | CoreDNS component monitoring | Managed Service for Prometheus | CoreDNS |
| Storage monitoring | Overview of container storage monitoring | Managed Service for Prometheus | csi-plugin, storage-operator |
| GPU monitoring | GPU monitoring | Managed Service for Prometheus | ack-gpu-exporter |
| Cluster topology monitoring | Cluster topology monitoring | Application Monitoring eBPF Edition | ack-arms-cmonitor |
| Cluster basic monitoring | Use Managed Service for Prometheus for monitoring | Managed Service for Prometheus | The Prometheus agent collects data from kube-state-metrics (Kubernetes object status), node-exporter (node metrics), kubelet (container metrics), and the ACK Virtual Node component. |
| Cost management suite | Enable cost insights | Managed Service for Prometheus | ack-cost-exporter |
| Cost management suite | Resource profiling | Managed Service for Prometheus | ack-koordinator (ack-slo-manager) |
| Scheduling | QoS aware scheduling | Managed Service for Prometheus | — |
| Scheduling | Best practices for colocation of multi-types workloads | Managed Service for Prometheus | — |
| Events and alerts monitoring | Event monitoring | Simple Log Service | ack-node-problem-detector |
| Events and alerts monitoring | Container service alert management | Simple Log Service, Managed Service for Prometheus, CloudMonitor | alicloud-monitor-controller |
| Kernel-level container monitoring | Kernel-level container monitoring based on SysOM | Managed Service for Prometheus | ack-sysom-monitor |
| Kernel-level container monitoring | Use SysOM to locate container memory issues | Managed Service for Prometheus | — |
| Network diagnosis | Network diagnosis | Managed Service for Prometheus | net-exporter |
| Backup center | Use Managed Service for Prometheus to monitor the backup center and configure alerting | Managed Service for Prometheus | migrate-controller |
| Knative | View the Knative monitoring dashboard | Managed Service for Prometheus | Knative |
| Auto Scaling | Horizontal pod scaling based on Managed Service for Prometheus metrics | Managed Service for Prometheus | ack-alibaba-cloud-metrics-adapter |
| Auto Scaling | Enable Managed Service for Prometheus for AHPA | Managed Service for Prometheus | AHPA controller |
| Cloud-native AI suite | Configure the Arena client | Managed Service for Prometheus | ack-arena |
| Cloud-native AI suite | Elastic datasets | Managed Service for Prometheus | ack-fluid |