Skip to content

Releases: opsrobot-ai/opsrobot

1.0.5

Choose a tag to compare

@littlestar1998 littlestar1998 released this 18 May 03:41
2e5e8b0

🚀 OpenClaw Observability Platform v1.0.5 Release Notes

[cite_start]The opsRobot 1.0.5 release delivers comprehensive, production-grade distributed tracing capabilities for OpenClaw[cite: 1, 6, 7]. [cite_start]Built on top of the KWeaver Core framework, this version leverages the OpenTelemetry (OTel) protocol and eBPF technology to provide full-link distributed tracing, deep bottleneck identification, and precise resource/cost tracking for AI Agents[cite: 3, 7].

🌟 Key Features

📊 1. Global Call Chain Dashboard (Control Room)

  • [cite_start]Visual Telemetry Summary: Direct visualization of core performance KPIs including total requests, success/failure counts, overall success rate, and latency metrics (average/maximum) with period-over-period trend analysis[cite: 13, 14].
  • [cite_start]Real-time Traffic Ingestion: Multi-dimensional filtering across multiple time horizons (ranging from the last hour up to 7 days) paired with granular 1-minute data aggregations[cite: 11, 12].
  • [cite_start]Anomaly & Bottleneck Identification: Instant ranking panels featuring top problematic endpoints sorted by highest latency, lowest success rates, and peak failure frequencies[cite: 15, 16].
  • [cite_start]Business-to-Infrastructure Correlation: Comprehensive views tracing request counts down to independent message channels alongside direct token consumption rankings grouped by LLM models[cite: 17].

🖥️ 2. Centralized Instance & Cluster Tracking

  • [cite_start]Massive Cluster Management: Centralized monitoring list built to track instance clusters with color-coded operational states: Normal (Green), Warning (Yellow), Critical (Red), and Offline (Grey)[cite: 20, 25].
  • [cite_start]Multi-Dimensional Querying: Fuzzy search capabilities matching against Instance ID, Hostname, or IP Addresses mixed with multi-condition environment filters[cite: 21, 23].
  • [cite_start]Deep Real-time Inspection: Granular execution statistics per instance highlighting total execution calls, failure rates, latency metrics, and high-latency markers[cite: 22, 24].

🔍 3. Single-Instance Micro-Diagnostics & Apdex Metrics

  • [cite_start]Advanced Experience Profiling: Support for user-defined Apdex (Application Performance Index) response thresholds ($T$) to quantify and graph satisfaction ratios and trace quality of service constraints[cite: 35, 81, 82, 83].
  • [cite_start]Scatter Plot Analytics: Scatter charts categorizing normal queries (blue) against failing or lagging queries (red) mapped over 5s/10s timeout thresholds[cite: 34, 73, 76].
  • [cite_start]Full-Stack Aggregation: Custom dimension cross-comparison across unique interface calls, applications, or host locations to identify transient performance fluctuations[cite: 36, 88, 89, 91].

🪵 4. Distributed Span Telemetry & Flame Graph Root-Cause Trace

  • [cite_start]Granular Span Audit Execution: Deep inspection logging for singular execution footprints tracking Parent Span IDs, interface definitions, execution states, and custom attribute tags[cite: 40, 44, 50, 54].
  • [cite_start]Flame Graph Diagnostics: Highly interactive hierarchical flame graph panels featuring synchronized node expansion, canvas zooming, and one-click raw error stack dumps for zeroing in on code exception sources[cite: 49, 53, 65, 68].
  • [cite_start]Trace-Level Cost Tracing: Contextual cost-attribution tracking that maps LLM token consumption and pricing tiers directly to individual trace legs (Session, Agent, and Model-specific costs)[cite: 66, 70].

🕸️ 5. Interactive Topology Engine

  • [cite_start]Automated Architecture Mapping: Dependency graphing engine automatically rendering inter-service downstream dependencies and request flow pathways[cite: 37, 96, 97].
  • [cite_start]Dynamic Node Sizing: Architectural nodes scale fluidly based on volumetric real-time query density, while localized dependency links turn bright red during active failures to map blast radiuses[cite: 37, 98, 100, 101].

📦 Quick Start & Resources

  • [cite_start]Source Code & Packages: Available on the official GitHub Repository[cite: 105, 107].
  • [cite_start]Container Deployment: Refer to the Quick Start Guide for standard Docker configurations[cite: 108, 109].
  • [cite_start]Issue Tracking & Support: Please report bugs or structural performance issues via GitHub Issues[cite: 108, 111].

1.0.4

Choose a tag to compare

@littlestar1998 littlestar1998 released this 09 May 09:43
2fabd55

opsRobot v1.0.4 Release Notes

This release introduces comprehensive Observability for OpenClaw Scheduled Tasks, providing administrators with a global cockpit to monitor, manage, and troubleshoot automated workflows.


🚀 New Features: OpenClaw Scheduled Task Observability

The core focus of v1.0.4 is the implementation of a full-stack monitoring solution for scheduled tasks, ensuring stability and performance across the platform.

Global Monitoring Dashboard (Running Overview):

  • Provides high-level visualization of task health, including total tasks, success rates, and failure counts.
    Token Consumption Monitoring: Track Token usage trends by date and type to optimize resource allocation.
    Diagnostic Heatmaps & Analytics: Identify execution fluctuations and high-frequency failure roots via trend curves and daily success rate heatmaps.
    Top 10 Performance Analysis: Rank tasks by execution frequency, failure rate, and latency to identify high-load or unstable operations.

Centralized Task Management (Task Monitor):

  • Monitor all scheduled tasks in a unified list with multi-dimensional filtering (Status, Agent, ID, etc.).
  • Real-time control over task activation/deactivation and execution strategy visibility.
  • Support for both Card and Table views to adapt to different operational scenarios.

Deep Traceability & Logging (Running Logs):

  • Centralized query interface for all execution records, supporting full-process auditing and problem tracing.

Log Detail Inspection: Expandable views for raw JSON outputs, execution summaries, and specific error messages.

Full-Link Traceability: Navigate from task execution points directly to Agent and Session details for rapid root-cause analysis.

Single-Task Deep Analytics:

  • Granular performance evaluation for individual tasks, including historical success rates and latency distribution.
  • Aggregated views of task configuration, execution logic, and step-by-step instructions.

1.0.3

Choose a tag to compare

@littlestar1998 littlestar1998 released this 27 Apr 01:27
364aaec

What's Changed

Full Changelog: 1.0.2...1.0.3

1.0.2

Choose a tag to compare

@littlestar1998 littlestar1998 released this 20 Apr 02:51
3b4e319

What's Changed

New Contributors

Full Changelog: 1.0.0...1.0.2

1.0.1-alpha

1.0.1-alpha Pre-release
Pre-release

Choose a tag to compare

@littlestar1998 littlestar1998 released this 13 Apr 03:40
61fcbc3

What's Changed

New Contributors

Full Changelog: 1.0.0...1.0.1-alpha

1.0.0

1.0.0 Pre-release
Pre-release

Choose a tag to compare

@littlestar1998 littlestar1998 released this 03 Apr 10:49
d5e636d

What's Changed

New Contributors

Full Changelog: https://github.com/opsrobot-observability/openclaw-observability-platform/commits/1.0.0