10 years of experience led to 100+ custom system checks for Apache RocketMQ, covering risks beyond regular monitoring like parameter tuning and internal status checks.
Systematic risk grading, orderly risk management
Smartly attributes and grades inspection risks, allowing users to manage risks by priority and root cause, ending disordered risk management.
Expert advice for fundamental risk elimination
Offers effective solutions for different risks, helping users eliminate risks at the root and prevent recurrence.
One-click diagnostic templates enable quick and easy pinpointing of past anomalies.
Built-in diagnostic templates for efficient self-determination of root causes
RocketMQ Copilot provides diagnostic templates for common issues like message backlog and un-consumption. A click identifies the root cause, easing troubleshooting and reducing emergency response time.
Flashback query to lock down problem scene
RocketMQ Copilot retains historical inspection anomaly information, making it easy to recreate the scene of occasional and difficult-to-reproduce problems.
Measure cluster stability digitally with SLI metrics, manage SLO, and handle ineffective alerts and alarm fatigue.
Built-in probes for end-to-end stability observation from an application perspective
Designed 12 E2E probes for RocketMQ, including message production success rate and RT, to detect any cluster abnormalities from an application's perspective.
Aim for SLO alerts, eliminate alarm fatigue
Aim for SLO alerts, support continuous and cumulative amount triggers, filter out invalid and scattered alarms to prevent alarm fatigue.
Automatically evaluates cluster performance baseline; intelligently predicts future water level usage and gives early warnings, calmly handling large-scale expansion and contraction.
Intelligently evaluate performance baselines and confirm the upper limit of cluster capacity
Automatically derive the cluster baseline based on machine specs and historical operation, guiding the capacity limit of the business cluster.
Predict usage in advance, calmly handle resource planning and expansion operations
Built-in predictive algorithm forecasts the usage for the next 30 days and sends alert notification 7 days in advance to prepare for resource scaling.
Lightweight and non-intrusive
Platform-agnostic, lightweight, non-intrusive deployment with no internet or external dependencies, seamlessly integrates with existing RocketMQ clusters.
Supports cross-platform installation and deployment on various architectures (x86, ARM, domestic CPUs) and OS (Linux, Windows, MacOS).
Run offline on a single 2c4G machine without internet or external service dependencies.
Non-intrusive to existing clusters.
Non-intrusive to existing RocketMQ clusters, compatible without any modifications.