Module analyze 

Deterministic analysis and insight engine.

Provides anomaly detection, pattern recognition, and diagnostic summaries for Docker containers based on metrics, events, and state snapshots. All analyses are fully deterministic — no AI/LLM calls are made.

Structs

AnalysisThresholds

Configurable thresholds for anomaly detection.

Anomaly

A single detected anomaly or insight.

ContainerExplanation

Summary of a single container’s health signals.

Signal

A diagnostic signal extracted from telemetry data.

Enums

AnalyzeError

Errors that can occur during analysis.

Severity

Severity level for a detected anomaly.

SignalStatus

Status classification for a signal.

Functions

analyze_density

Analyze container density / distribution across images, states, and compose projects.

analyze_dependencies

Analyze container dependencies: which containers share networks, compose projects, and images.

detect_config_drift

Detect configuration drift between historical snapshots.

detect_deployment_errors

Detect deployment-related error spikes from historical events.

detect_high_cpu

Detect high CPU usage anomalies.

detect_high_disk_io

Detect high disk I/O (read/write) above thresholds.

detect_high_network_io

Detect high network I/O (RX/TX) above thresholds.

detect_memory_pressure

Detect memory pressure: containers with memory usage/limit ratio above threshold.

detect_resource_leaks

Detect containers whose memory usage is trending upward over time, indicating a possible memory leak.

detect_restart_loops

Detect restart loops: containers with restart_count >= threshold.

detect_unhealthy_states

Detect containers in unhealthy states (exited, dead, restarting).

execute_analyze

Execute an analysis query using live Docker data + metrics. Thresholds are loaded from the user’s config file (if available), falling back to defaults.

execute_analyze_with_store

Execute an analysis query using historical data from a telemetry store. Thresholds are loaded from the user’s config file (if available), falling back to defaults.

execute_analyze_with_thresholds

Execute an analysis query with custom thresholds.