Cloud Monitor:What is Managed Service for Prometheus?

1. Data collection

Alibaba Cloud Managed Service for Prometheus provides full-stack metric collection for client-side applications, server-side applications, containers, hosts, self-managed components, and Alibaba Cloud services. It also supports integrating data from your self-managed Prometheus environments using the standard open source Prometheus Remote Write protocol. You can also write custom metrics using a software development kit (SDK).

• Client-side & applications: Metrics from ARMS Real User Monitoring (RUM) and ARMS Application Performance Management (APM) are written to Prometheus by default. For custom metrics in your applications, you can write them using the Prometheus SDK or the OpenTelemetry SDK.

• Containers: Alibaba Cloud container services, such as ACK, ASK, and ACS, integrate Managed Service for Prometheus by default. A self-developed Prometheus Agent collects basic container monitoring metrics. You can also collect custom metrics using methods such as ServiceMonitor and PodMonitor. For self-managed Kubernetes clusters or Kubernetes services from other cloud providers, you can use ACK One to manage them as registered clusters. This provides the same container metric collection capabilities as ACK.

• Hosts: For virtual host monitoring, Managed Service for Prometheus uses a managed Prometheus Agent to collect operating system metrics, such as CPU, memory, disk, and network usage. It also collects metrics related to processes and containers. On ECS, you can collect custom metrics using a textfile method similar to that of node-exporter. For virtual hosts in self-managed data centers or from other cloud providers, you can collect data using the open source node-exporter and Prometheus components and then use the Remote Write protocol to centralize the data in Managed Service for Prometheus.

• Alibaba Cloud services: Managed Service for Prometheus supports data integration from Alibaba Cloud Monitor through a simple user interface. This lets you collect monitoring metrics for Alibaba Cloud services. The cloud service integration feature automatically enriches these metrics with information, such as instance names and tags. This provides more dimensions for aggregation, filtering, and alert routing.

• Custom metrics: Managed Service for Prometheus is compatible with the open source Prometheus and OpenTelemetry metric write protocols. You can use an SDK to write custom metrics.

2. Data storage

Alibaba Cloud Managed Service for Prometheus supports tiered data storage. After data in Standard storage expires, it is automatically transferred to Archive Storage. This meets your needs for long-term metric storage at a low cost.

• Managed Service for Prometheus Standard storage offers different storage durations. Billing is based on the volume of data written or reported.

• If you need a longer storage duration, you can enable Archive Storage. After you enable this feature, data in Standard storage is automatically moved to Archive Storage when it expires. Archive Storage is billed based on the amount of data stored.

3. Data consumption

Alibaba Cloud Managed Service for Prometheus provides unified metric collection and storage to support various data consumption scenarios.

• Unified visualization: Managed Service for Prometheus offers a rich set of out-of-the-box dashboards for Alibaba Cloud services and common open source components. You can import these built-in dashboards into Managed Service for Grafana or create custom dashboards. You can also connect to Alibaba Cloud's DataV platform through the Prometheus HTTP API to build more dynamic visualizations.

• Aggregated views: The aggregated view feature in Managed Service for Prometheus supports aggregate queries across instances and accounts. This enables unified monitoring across different accounts.

• Data processing: Managed Service for Prometheus provides pre-aggregation using Recording Rules. It is compatible with open source Recording Rules and lets you perform downsampling and dimension reduction on metrics. This reduces data costs and improves query performance.

• Unified alerting & response: Managed Service for Prometheus includes default alert rules for Alibaba Cloud services and common open source components. You can also define custom alert rules using PromQL or import alerting rules from open source Prometheus. When an alert rule is triggered, an event is sent to the ARMS alert management system, which then handles actions such as notifications, assignments, and escalations.

• CloudLens: For core monitoring scenarios, CloudLens provides unified observability based on logs, metrics, and events. This provides a complete view of your Alibaba Cloud services.

• Data analytics: Managed Service for Prometheus is 100% compatible with PromQL syntax. You can use PromQL to perform custom analysis and exploration. The service also provides metric management capabilities that help you understand metric distribution and identify high-cardinality situations.

• Data exporting: Managed Service for Prometheus supports real-time data delivery. You can deliver metric data to Kafka, MaxCompute, or a self-managed Prometheus instance.