Features and benefits-Scalable Open Financial Architecture(SOFAStack)-阿里云帮助中心

This topic describes the features of Real-time Monitoring Service (RMS). Before you use this service, we recommend that you learn the functional architecture and basic concepts of RMS to better understand RMS features.

Risk discovery

The following table describes the RMS features related to risk discovery.

Category	Feature	Description
Monitoring	Business monitoring	Allows you to flexibly split business logs of applications. Aggregates data by business dimensions and provides a comprehensive set of capabilities to support real-time business analysis. Applicable to personalized and non-standard business monitoring scenarios, such as transaction volume monitoring and loan monitoring. Collects data of key metrics within seconds and common metrics within minutes in a highly reliable way.
	Application monitoring	Supports the monitoring of both container applications and classic applications. Monitors an application inside out based on multi-dimensional metrics, such as service metrics, system metrics, Java virtual machine (JVM) metrics, and port detection metrics Aggregates metrics from different perspectives such as the single-instance perspective, IDC perspective, and LDC perspective.
	Host monitoring	Supports the monitoring of physical machines and VMs based on various metrics such as system metrics and port detection metrics.
	Cloud-native monitoring	Provides bountiful capabilities for monitoring cloud-native resources such as Kubernetes clusters, Kubernetes components, and container runtimes. By default, a visual panel is provided for you to monitor cloud-native resources in an optimal way. Allows you to make comparisons and perform analysis from various perspectives such as the multi-cluster perspective, single-cluster perspective, and application perspective. This helps O&M engineers better manage the PaaS platform.
	Prometheus monitoring	Collects Prometheus metrics that meet the requirements of OpenMetrics, stores metric data as time series data in a unified format, and provides a dashboard for you to view the metrics.
	Monitoring of Ant Group services	Collects native metrics from Scalable Open Financial Architecture (SOFA)-based logs. By default, this feature monitors Ant Group services and allows you to configure alert rules for these services.
	Self-monitoring	Monitors and inspects the availability of RMS components, and allows you to deploy RMS-SM to comprehensively monitor RMS.
Monitoring	One-stop application analysis	Implements one-stop monitoring on applications based on metadata. Allows you to monitor an application based on multi-dimensional metrics such as service metrics, system metrics, JVM metrics, and port detection metrics. Aggregates metrics from different perspectives such as the single-instance perspective, IDC perspective, and LDC perspective, and performs multi-dimensional correlation analysis on applications, services, instances, and hosts throughout the roll-up and drill-down processes. This helps you troubleshoot errors with ease.
Alerting	Alert rules	Allows you to flexibly configure alert rules for monitored objects. If an alert rule is triggered, RMS generates an alert event. For example, RMS sends alert notifications to alert contacts to inform them of the situation and ask them to take necessary measures.
Alerting	Alert templates	Allows you to use an alert template to create multiple alert rules at a time. This improves the efficiency of configuring alert rules.
	Alert rules	Allows you to flexibly configure alert rules for monitored objects. If an alert rule is triggered, RMS generates an alert event. For example, RMS sends alert notifications to alert contacts to inform them of the situation and ask them to take necessary measures.
	Alert events	An alert event refers to the full-lifecycle management of an alert from when the alert is generated to when the alert is cleared. Each alert event has a unique ID. If an alert rule is triggered, an alert is generated. All alerts generated after the alert rule is triggered and before the alert status is cleared belong to the same event.
	Alert clearance	Alert clearance is an anti-jitter mechanism for alerts. If an alert rule is triggered during a specific number (N) of consecutive detection cycles, an alert record is generated. Then, if the alert rule is not triggered during N consecutive detection cycles, an alert clearance record is generated.
	Alert silence	If an alert rule keeps being triggered, RMS sends an alert notification after the alert silence period elapses. If the monitored data enters a normal state, a notification is sent to inform each relevant contact of the normal status and the alert is cleared. If the monitored data triggers an alert rule again, an alert notification is sent to each relevant contact.
	Alert pause	You can enable the alert pause feature for an alert rule for specific reasons. During the alert pause period, no detection is carried out to check whether the alert rule is triggered and no alert is generated.
Template-based configuration	Template-based configuration	Allows you to export application monitoring and custom monitoring configurations as templates in the JSON format. Then, you can import these templates to other environments to quickly reuse the configurations. If an application with the specified name has been deployed in an environment to which a monitoring template is imported, the alert configuration in the template applies to the application. This makes production and O&M more efficient, convenient, and consistent.

Risk positioning

The following table describes the RMS features related to risk positioning.

Category	Feature	Description
Tracing analysis	Global application topology	Continuously and automatically discovers inter-application calls, constructs a global topology based on the dependencies and call relationships between middleware services, and displays the exception status of nodes and lines in the topology. This way, you can be aware of failures and problems at the earliest opportunity.
	Application drill-down and performance analysis	Supports drill-down analysis layer by layer from base applications to upstream and downstream applications and to APIs, displays key performance metrics such as the throughput, average response time, error rate, and satisfaction, and establishes data associations from the underlying layer to the upper layers. This helps you troubleshoot the root problems that affect application performance in distributed scenarios.
	Multi-dimensional trace query	Allows you to query traces based on multi-dimensional business conditions. For example, you can query traces by using different call methods such as remote procedure calls (RPCs), the model-view-controller (MVC), RESTful APIs, and databases. You can customize tags such as transaction numbers based on scenarios to query the corresponding traces.
	Trace details	Provides trace diagrams, trace details, and sequence diagrams for inter-system calls. This clarifies the interaction, performance, and exceptions of inter-system calls in an all-around way.
	Compatibility with OpenTelemetry	Seamlessly integrates SOFAStack applications with OpenTelemetry. You can import your business code to RMS without modifications. RMS also allows you to import trace data from Zipkin or Apache SkyWalking, which complies with OpenTracing requirements, to generate application topologies and traces for tracing analysis.
Tracing analysis	Trace-log association	Supports various log data sources and allows you to associate traces with business logs to increase the business information in traces.
Logging	Log query	Supports exact queries, fuzzy queries, full-text queries, keyword queries, and log context queries.
	LiveTail	In online O&M scenarios, you often need to monitor the data in the log queue in real time and extract key information from the latest log data to quickly identify the cause of an exception. LiveTail that is provided by RMS allows you to monitor and analyze online log data in real time. This simplifies the O&M process.
	Log snapshots	Allows you to save frequently used query conditions and import the saved query conditions with a few clicks.
	Log download	Allows you to download and export logs that are generated within the specified time range in CSV or TXT format.
	Log collection	Allows you to collect logs of containerized applications and classic applications. You can specify a log blacklist or whitelist during log collection configuration, and collect and process log content based on the log collection configuration.
	Log dump	RMS-LS supports the log dump feature. To store logs for an extended period of time or back up logs, you must store or write the logs to disks.
	Index management	Supports comprehensive management throughout the lifecycle of index storage. This helps you better manage logs to ensure the high availability of log data. You can store logs in each workspace by using an index and configure the storage duration of logs.

Benefits

RMS has the following benefits:

Low resource consumption
RMS can reliably collect and transmit a large amount of data but consume only a small number of host resources, such as CPU and memory resources.
Stable and efficient data computing and storage
RMS can monitor tens of thousands of devices in real time and collect metric data within minutes, and can continuously aggregate data online. In addition, RMS intelligently classifies and stores data, controls data volumes in time series databases, and allows you to perform elastic scaling in the RMS console.
High availability
RMS supports disaster recovery by using two IDCs, runs components on multiple instances, and synchronizes databases by using multiple replicas. In addition, RMS can monitor its components and supports automatic fault recovery.
Open and independent features
All RMS components are self-developed and controllable, and have been verified by massive experiments. RMS provides comprehensive features and data APIs to better integrate with third-party systems.