Docs

Observability Platform

What's new

Release notes

This topic describes the latest updates for the feature modules related to Observability Platform. New features will be rolled out across regions. We are looking forward to your hands-on experience.

tip

The Regions column lists the regions where the feature change applies at the time of the release. The actual supported regions are subject to those displayed in the console. Supported regions vary by module. All indicates all regions where the current module is available.

September 2026

Module	Feature	Description	Regions	Related document
AI application monitoring	AI Application Insights	This feature provides intelligent analysis capabilities for session data. You can perform batch analysis on sessions within a specified range using preset insight rules. This applies to scenarios such as session satisfaction analysis, task completion assessment, and exception session filtering.	All	AI Application Insights
Integration center	Integrating AI applications	Integration center provides multiple AI application integration solutions, supporting one-click generation of installation commands based on basic parameters. This simplifies AI application integration and enables rapid construction of an end-to-end trace observability system for AI applications. Currently supported AI applications include: Codex, Claude, Cursor, OpenCode, OpenClaw, DeepSeek Harness, and Hermes.	All	Integrating AI applications
Integration center	Host probes	Probe management now supports installing and managing host probes.	All	Probe management
Integration center	Probe management	A dedicated probe installation page is provided to facilitate quick installation of Volcano Engine VKE probes, third-party cloud container probes, and host probes.	All	Probe management
Integration center	Integrating server-side applications from third-party clouds	Multi-cloud integration now supports integrating server-side applications from AWS and self-managed Kubernetes into Server-side Monitoring for data collection and probe management.	All	Integrating server-side applications from third-party clouds

August 2026

Module	Feature	Description	Regions	Related document
AI application monitoring	LLM trace analysis	Large model call traces in AI applications can now be filtered and annotated with user feedback. Users can rate individual traces with a thumbs up or thumbs down, enabling continuous model performance evaluation and issue attribution based on user feedback.	All	AI Application Trace Analysis
AI application monitoring	Playground	In Playground, multi-turn conversations can now be built, prompts and model parameters can be adjusted, tool calls can be simulated, and output formats can be constrained, with immediate execution and result preview. It is suitable for online bad case review, prompt tuning, model selection, parameter tuning, tool call debugging, and structured output validation.	All	Playground
Integration center	Multi-cloud integration task and probe management	Multi-cloud integration is available for connecting server-side applications on Tencent Cloud, Huawei Cloud, or Alibaba Cloud to Server-side Monitoring for data collection and probe management.	All	Integrating server-side applications from third-party clouds Probe management
Integration center	Probe configuration updates	Deployment configurations for probes in VKE environments can now be updated, including deployment method, number of instances, CPU requests/limits, and memory requests/limits.	All	Probe management
Alert Center	Unified management of alert configurations	The Alert Center of Observability Platform consolidates alert configurations across all observability products for unified management. This helps the platform continuously improve alert governance, rule auditing, policy optimization, and alert operations, enhancing the maintainability and scalability of the overall observability system. Observability products include Cloud Monitor, cloud service events, VMP, and Server-side Monitoring. Consolidated alert configurations include alert rules, event rules, alert templates, preset alert templates, and object groups.	All	Cloud Monitor alert configurations Alert configurations for cloud service events Alert configurations for VMP Alert configurations for Server-side Monitoring
Alert Center	Import/export alert rules	Alert rules for Cloud Monitor, cloud service events, VMP, and Server-side Monitoring can now be imported and exported, enabling quick synchronization of alert configurations across accounts, regions, projects, or workspaces for efficient reuse.	All	Managing Cloud Monitor alert rules Managing cloud service event rules Managing VMP alert rules Managing Server-side Monitoring alert rules
Alert Center	Import/export alert templates	Alert templates for Cloud Monitor, VMP, and Server-side Monitoring can now be imported and exported, enabling quick synchronization of alert configurations across accounts, regions, projects, or workspaces for efficient reuse.	All	Managing Cloud Monitor alert templates Managing VMP alert templates Managing Server-side Monitoring alert templates
Alert Center	Alert overview	The alert overview aggregates alert data across the full product lines of Cloud Monitor, VMP, and Server-side Monitoring, providing a clear view of overall failure risk and alert sources, and enabling identification of alert surge points and change patterns through alert trend analysis.	All	Alert overview
Alert Center	Alert event	When an alert is triggered, the system records alert event information such as alert status, trigger conditions, and alert time to facilitate traceability and troubleshooting.	All	Alert event
Alert Center	Notification history	When an alert is triggered, the system records notification history information such as message content, notification time, and delivery results, as well as alert suppression information to facilitate traceability and troubleshooting.	All	Notification history
Alert Center	Aggregate alert events by contact	Notification policies can now aggregate alert events by contact and send them as a single notification, significantly reducing notification volume. This capability is suitable for alert events that are low in priority but high in volume.	All	Creating notification policies

July 2026

Module	Feature	Description	Regions	Related document
Intelligent O&M	Automation	Automated tasks have been introduced, enabling observability tasks to be executed on schedule. This feature is suitable for periodic use cases such as routine inspections and resource health checks.	All	Configuring automated tasks
Intelligent O&M	Expert agent and skill	Skills and agents derived from observability experience can now be configured in the intelligent O&M assistant to extend its capabilities.	All	Custom expert agent Skill management
Integration center	Integration Center official release	Integration center is the management center on the Observability Platform for collecting multi-source, heterogeneous metrics, logs, and traces. It provides visual, out-of-the-box capabilities such as data collection configuration, collection task management, and collection probe management, significantly lowering the barrier to data collection and observability. In the Integration center first release, server-side application integration, integration task management, and probe management are provided.	All	Integrating server-side applications in VKE Integration management Probe management
Metric center	Metric management	Metric management is a visual metric search page provided by Observability Platform. It allows you to quickly search metrics by product, time, metric type, attribute category, and more.	All	Metric management
Intelligent O&M	VeOps CLI observability troubleshooting assistant	VeOps CLI is an observability troubleshooting tool designed for AI agents. It wraps the capabilities of Server-side Monitoring, VMP, Cloud Monitor, TLS, and VKE, as well as local log and metric analysis, into commands that general-purpose agents can reliably call.	All	VeOps CLI observability troubleshooting assistant
Custom dashboard	Data transformation	Added five data transformation policies: Add field from calculation, Reduce, Rename fields by regex, Group by, and Series to rows.	All	Data transformation
Custom dashboard	Version management	Added the version management feature, allowing you to create, compare, and restore versions of custom dashboards. The feature records every change and standardizes the update process for custom dashboards, preventing collaborators from overwriting each other's work.	All	Version management
Alert Center	Alert Center official release	The Alert Center is a unified alert management module provided by Observability Platform for all observability products. It not only normalizes alert and notification configuration data, but also provides a centralized entry point for managing alert rules. This breaks down alert configuration silos across observability products and lays the foundation for improving the standardization and maintainability of the overall observability system.	All	Alert Center overview

June 2026

Module	Feature	Description	Regions	Related document
Intelligent O&M assistant	Public beta launch of Intelligent Operations Assistant	[Public beta] The intelligent O&M assistant is available in public beta for the first time. Built on monitoring data and designed for AI O&M, it serves as a full-stack intelligent O&M hub. Primarily tailored for observability scenarios, the intelligent O&M assistant possesses in-depth domain expertise and provides data query, basic analysis, and cross-service comprehensive troubleshooting capabilities.	All	Overview of intelligent O&M assistant Using intelligent O&M assistant
AgentKit application observability	Insight dashboard	The AgentKit insight dashboard is optimized, and monitoring dashboards for key metrics of usage analysis and performance analysis are now available.	All	Insight dashboards
AI application monitoring	Diagnose a session	Session diagnosis is provided for AI applications. Trace and Session data of the current session can be temporarily used for diagnosis and analysis, and detailed diagnosis reports are provided.	All	LLM session analysis Observability diagnosis
AgentKit application observability	Diagnose a session	Session diagnosis is provided for AgentKit applications. Trace and Session data of the current session can be temporarily used for diagnosis and analysis, and detailed diagnosis reports are provided.	All	Session analysis Observability diagnosis
Server-side Monitoring	AI agent application integration	OpenClaw, Hermes, and Claude Code are integrated into APMPlus via plugins to implement trace monitoring.	All	OpenClaw trace monitoring Hermes trace monitoring Claude Code trace monitoring
AI application monitoring	Export of session call trace data	Detailed trace information of target sessions can be exported in LLM session analysis and AgentKit session analysis.	All	LLM session analysis AgentKit session analysis
Custom dashboard	Bar charts and text charts	Bar charts and text charts are newly added to custom dashboards to meet usage requirements in various scenarios.	All	Chart settings
Custom dashboard	Display names configurable for time series charts	Display names can be configured for metrics in the "Overrides" settings of time series charts.	All	Chart settings
AI application monitoring	Session data export	Exporting session data based on actual requirements is now supported in LLM and AgentKit session analysis.	All	LLM session analysis AgentKit session analysis

May 2026

Module	Feature	Description	Regions	Related document
General	Root Span list	A Root Span list is displayed in the Trace analysis module. All Root Span-related Traces and Trace-related details within the specified filter range are displayed.	All	Server-side Monitoring Trace Analysis AI Application Trace Analysis AgentKit trace analysis

March 2026

Module	Feature	Description	Regions	Related document
AI application monitoring	Trace diagnosis	Intelligently analyzes an AI trace via AgentLens and generates a detailed diagnostic report. The report covers core runtime data statistics, issue summary, input/output overview, execution process analysis, and in-depth diagnosis results.	All	General AI trace analysis AgentKit trace analysis
AI application monitoring	Online trace export	By exporting trace data in batches to an evaluation set on the agent platform, you can run evaluation experiments directly using trace samples from the evaluation set. This process forms a closed loop from real online data to evaluation and verification, improving the accuracy and reliability of diagnostic capabilities on the observability platform.	China (Beijing) China (Shanghai)	Online trace export
AgentKit application monitoring	Insight dashboard now supports session dashboards.	Analysis of all sessions within AgentKit: visualized dashboards of session-related data and trends. Currently supported: trend dashboards for active session count, average session event count, and average session token consumption count.	All	Insight dashboards

February 2026

Module	Feature	Description	Regions	Related document
Server-side Monitoring	OpenClaw observability	APMPlus is the first to support receiving data reported by the diagnostics-otel plugin and provides a ready-to-use OpenClaw monitoring dashboard.	All	OpenClaw trace monitoring
Custom dashboard	Pie chart and bar chart	Added support of pie chart and bar chart in chart settings to meet the needs of different scenarios.	All	Chart settings
Custom dashboard	Metrics data source	Data query now supports the "Metrics" data source. When "Metrics" is selected as the data source, metric data can be queried via OpenTSDB and InfluxDB protocols.	All	Data query
Custom dashboard	Data link	Time series diagrams support data link configuration, enabling quick navigation to specified pages.	All	Link configuration

January 2026

Module	Feature	Description	Regions	Related document
AgentKit application monitoring	Preset alert templates	The system provides preset alert templates for resources in AgentKit, such as agent runtimes and tools. This enables you to create alert tasks quickly by using preset alert templates and monitor the health status of your services in real time.	All	Alert management
AI application monitoring	Multimodal content display	For multimodal scenarios, switch to the LLM data tab on the single trace details page to view the Markdown-formatted text and multimodal content from the Input and Output fields.	All	Trace analysis
AgentKit application monitoring	Tool analysis	Data detail monitoring is available for sandbox tools. This enables viewing key metrics such as tool call count, error rate, and average response time.	All	Tool analysis
AgentKit application monitoring	Runtime monitoring	The Runtime details feature offers monitoring dashboards that allow you to view the resource consumption of the current runtime within a specified time period. The information includes CPU duration, CPU usage, memory usage, number of GC objects, number of threads, and number of context switches.	All	Runtime monitoring

December 2025

Module	Feature	Description	Regions	Related document
Custom dashboard	Preset dashboards	Preconfigured observability dashboards for specific scenarios deliver out-of-the-box observability capabilities.	All	Preset dashboards
AI application monitoring	Display of tool information in trace analysis	Tool-related information can be displayed in the LLM data section of the AI-related trace analysis page.	All	Trace analysis
AgentKit application monitoring	Insight dashboards	A separate insight dashboard is provided for AgentKit. It supports the overview of AgentKit as well as the statistical and trend analysis of agent-related data, and facilitates health checks, performance optimization, anomaly detection, and other operational tasks.	All	Insight dashboards
AgentKit application monitoring	Knowledge base analysis	Knowledge base analysis is added. This supports the monitoring of knowledge detection data, such as API call count, average latency, and total error rate.	All	Knowledge base analysis
AgentKit application monitoring	Memory analysis	Memory analysis is supported for Viking-type memories.	All	Memory analysis

November 2025

Module	Feature	Description	Regions	Related document
AI application monitoring	LLM trace analysis	An independent trace analysis page is provided for general AI application monitoring. This facilitates fault troubleshooting, performance optimization, and system monitoring, and improves O&M efficiency.	All	LLM trace analysis
AI application monitoring	Optimized display of quantile charts	P99/P90/P75/P50 metric filtering is added for the charts of model request latency quantiles, TTFT quantiles, and TPOT quantiles.	All	LLM monitoring dashboard
General	First release of Observability Platform	Observability Platform is released for the first time. It aggregates multiple independent services, such as AI application monitoring, Server-side Monitoring, Volcengine Managed Service for Prometheus (VMP), Cloud Monitor, RUM, and CloudDetect. It also integrates universal modules, such as custom dashboards, an alert center, and an integration center, to deliver a streamlined, one-stop observability portal.	All	None

Last updated: 2026.09.21 10:36:58