Observability Event Types

Complete reference for system and conversation events

Ask ChatGPT Ask Claude.ai Open in Cursor

MUXI emits 350+ typed events across the request lifecycle. Route them to your logging infrastructure for debugging, auditing, and performance monitoring. Events are emitted as JSONL and can be filtered by pattern.

Event Envelope

Every event follows this structure:

{
  "id": "evt_abc123",
  "timestamp": 1706616000000,
  "level": "info",
  "event": "request.completed",
  "request": {
    "id": "req_xyz789",
    "status": "completed",
    "duration_ms": 842,
    "formation_id": "my-formation",
    "user_id": "usr_123",
    "session_id": "sess_456"
  },
  "data": {
    "agent": "researcher",
    "tokens_prompt": 1200,
    "tokens_completion": 450
  }
}

Request Ingestion & Validation

Events related to initial request processing, authentication, rate limiting, and validation.

Event	Level	Description
`request.received`	INFO	Request initially received by MUXI
`request.denied.auth`	WARN	Request denied due to authentication failure
`request.denied.rate_limit`	WARN	Request denied due to rate limiting
`request.denied.validation`	WARN	Request denied due to validation failure
`request.validated`	DEBUG	Request format and structure validation passed

Multi-Modal Content Processing

Events for processing documents, images, audio, and other media content types.

Event	Level	Description
`content.document.parsed`	INFO	Document attachment processing
`content.image.analyzed`	INFO	Image analysis and vision processing
`content.audio.transcribed`	INFO	Audio transcription
`content.extraction.failed`	ERROR	Content processing failure

Overlord Orchestration

Events tracking the overlord's routing decisions, task decomposition, and orchestration activities.

Event	Level	Description
`overlord.initialized`	DEBUG	Overlord component startup
`overlord.routing.started`	DEBUG	Begin routing decision process
`overlord.routing.completed`	INFO	Routing decision made
`overlord.task.decomposed`	INFO	Request broken into subtasks

Memory & Context Operations

Events for memory storage, retrieval, and context management across short-term and long-term memory systems.

Event	Level	Description
`memory.retrieval.started`	DEBUG	Memory search initiated
`memory.retrieval.short_term`	DEBUG	Short-term buffer search
`memory.retrieval.long_term`	DEBUG	Long-term memory search
`memory.retrieval.context`	DEBUG	User context memory lookup
`memory.storage.short_term`	DEBUG	Buffer memory storage
`memory.storage.long_term`	DEBUG	Persistent memory storage
`memory.extraction.started`	INFO	Automatic user info extraction

Agent Processing

Events tracking individual agent activities including reasoning, planning, and context application.

Event	Level	Description
`agent.selected`	INFO	Specific agent chosen for processing
`agent.thinking.started`	DEBUG	Agent reasoning process begins
`agent.thinking.completed`	INFO	Agent reasoning finished
`agent.planning.created`	INFO	Agent generates execution plan
`agent.context.applied`	DEBUG	System message/soul applied

Model Operations

Events for LLM API calls, inference operations, and streaming response handling.

Event	Level	Description
`model.inference.started`	DEBUG	Model API call initiated
`model.inference.completed`	INFO	Model response received
`model.streaming.started`	DEBUG	Streaming response initiated

Tool & MCP Operations

Events for MCP server connections, tool discovery, execution, and error handling.

Event	Level	Description
`mcp.connection.established`	DEBUG	MCP server connection
`mcp.tool.discovered`	DEBUG	Available tools identified
`mcp.tool.called`	INFO	Tool execution initiated
`mcp.tool.completed`	INFO	Tool execution finished
`mcp.tool.failed`	ERROR	Tool execution error

A2A & Collaboration

Events for agent-to-agent communication, external agent discovery, and multi-agent collaboration.

Event	Level	Description
`a2a.discovery.started`	DEBUG	External agent discovery
`a2a.request.sent`	INFO	A2A request to external agent
`a2a.response.received`	INFO	A2A response received
`collaboration.internal.started`	INFO	Internal multi-agent collaboration

Response Generation

Events for response composition, multi-modal content creation, validation, and formatting.

Event	Level	Description
`response.generation.started`	DEBUG	Begin response composition
`response.multimodal.created`	INFO	Multi-modal response generated
`response.validation.completed`	DEBUG	Response safety/validation check
`response.formatted`	DEBUG	Final response formatting

Async & Delivery

Events for asynchronous processing, webhook delivery, and response completion tracking.

Event	Level	Description
`async.threshold.detected`	INFO	Request exceeds sync threshold
`async.processing.started`	INFO	Background processing initiated
`webhook.sent`	INFO	Webhook delivery attempted
`response.delivered`	INFO	Final response sent to user

Error Handling & Recovery

Events for error detection, retry mechanisms, fallback activation, and recovery completion.

Event	Level	Description
`error.timeout.detected`	WARN	Operation timeout
`error.retry.attempted`	WARN	Retry mechanism triggered
`error.fallback.activated`	WARN	Fallback mechanism used
`error.recovery.completed`	INFO	Error recovery successful

Performance & Monitoring

Events for performance metrics, resource usage tracking, and session management.

Event	Level	Description
`performance.duration.recorded`	DEBUG	Component performance metric
`resource.usage.measured`	DEBUG	Resource consumption tracking
`session.created`	INFO	User session established
`session.context.updated`	DEBUG	Session context modified

Clarification & Parameter Collection

Events for intelligent parameter collection, clarifying questions, and missing parameter detection.

Event	Level	Description
`clarification.analysis.started`	DEBUG	Information requirement analysis initiated
`clarification.analysis.completed`	INFO	Missing parameters/information identified
`clarification.question.generated`	INFO	Clarifying question created for user
`clarification.response.parsed`	DEBUG	User clarification response processed
`clarification.parameter.enriched`	DEBUG	Parameter filled from user context
`clarification.parameter.validated`	DEBUG	Parameter validation completed
`clarification.failed`	WARN	Clarification process failed, falling back

Proactive Clarification & Modes

Events for proactive clarification sessions, user-requested questioning, and mode management.

Event	Level	Description
`clarification.proactive.detected`	INFO	Explicit turn-taking request detected
`clarification.mode.started`	INFO	Proactive clarification session initiated
`clarification.mode.questioning`	DEBUG	Proactive questioning mode active
`clarification.mode.planning`	DEBUG	Plan analysis mode active
`clarification.mode.context_building`	DEBUG	Context building mode active
`clarification.mode.goal_achievement`	DEBUG	Goal achievement mode active
`clarification.session.completed`	INFO	Proactive session reached completion criteria
`clarification.session.cancelled`	INFO	User cancelled proactive session

Planning Workflow Detection

Events for implicit planning workflow identification, data synthesis, and continuation.

Event	Level	Description
`planning.workflow.detected`	INFO	Implicit planning workflow identified
`planning.data.synthesized`	INFO	Tool results synthesized for decision-making
`planning.continuation.started`	INFO	Planning continuation after data gathering
`planning.step.analyzed`	DEBUG	Individual plan step feasibility assessed
`planning.dependencies.identified`	DEBUG	Plan step dependencies mapped
`planning.options.presented`	INFO	Decision options generated for user

Enhanced Tool Processing

Events for tool parameter collection, validation, and context-enhanced execution.

Event	Level	Description
`tool.parameter.missing`	INFO	Missing required tool parameters identified
`tool.parameter.clarified`	INFO	Tool parameter obtained via clarification
`tool.validation.failed`	WARN	Tool parameter validation failed
`tool.execution.enhanced`	DEBUG	Tool executed with enhanced parameter processing
`tool.context.applied`	DEBUG	User context applied to tool parameters

Multilingual & Detection

Events for language detection, multilingual processing, and pattern recognition.

Event	Level	Description
`language.detection.started`	DEBUG	LLM-based language detection initiated
`language.detection.completed`	DEBUG	Input language identified
`language.processing.multilingual`	DEBUG	Multilingual processing mode activated
`language.fallback.activated`	WARN	Fallback to regex patterns
`intent.proactive.detected`	INFO	Proactive clarification intent identified
`intent.planning.detected`	INFO	Planning workflow intent identified
`intent.goal.extracted`	DEBUG	User goal extracted from natural language
`pattern.llm.analysis`	DEBUG	LLM-based pattern analysis completed
`pattern.detection.failed`	WARN	Pattern detection failed, using fallback

Context & Information Management

Events for information requirements analysis, gap identification, and context management.

Event	Level	Description
`info.requirements.analyzed`	DEBUG	Information requirements determined
`info.gaps.identified`	INFO	Information gaps found in user request
`info.confidence.scored`	DEBUG	Information confidence level calculated
`info.extraction.automated`	DEBUG	Automatic information extraction attempted
`info.context.insufficient`	WARN	Insufficient context for parameter filling
`goal.type.determined`	DEBUG	Goal type classification completed
`goal.progress.tracked`	DEBUG	Progress toward goal completion measured
`goal.completion.criteria`	DEBUG	Goal completion criteria evaluated
`goal.context.created`	INFO	Goal context established for session
`goal.achievement.measured`	DEBUG	Goal achievement percentage calculated

Session & State Management

Events for clarification session lifecycle, state management, and mode transitions.

Event	Level	Description
`session.clarification.created`	INFO	New clarification session established
`session.clarification.updated`	DEBUG	Session state updated with new information
`session.clarification.cleaned`	DEBUG	Session cleanup and resource release
`session.mode.switched`	INFO	Clarification mode changed within session
`session.progress.checkpoint`	DEBUG	Session progress checkpoint reached
`conversation.turn.clarification`	DEBUG	Clarification turn in multi-turn conversation
`conversation.context.maintained`	DEBUG	Context preserved across clarification turns
`conversation.flow.interrupted`	WARN	Normal conversation flow interrupted for clarification
`conversation.flow.resumed`	INFO	Normal conversation flow resumed after clarification

Clarification Performance & Quality

Events for performance metrics, quality tracking, and system health monitoring.

Event	Level	Description
`performance.clarification.duration`	DEBUG	Time spent in clarification process
`performance.question.generation`	DEBUG	Question generation latency measured
`performance.llm.detection`	DEBUG	LLM-based detection performance
`quality.parameter.completion`	INFO	Parameter completion rate measured
`quality.clarification.success`	INFO	Clarification success rate tracked
`health.clarification.enabled`	INFO	Clarification system initialized successfully
`health.clarification.disabled`	WARN	Clarification system disabled/failed
`health.mode.manager.active`	DEBUG	Mode manager health check passed
`health.session.cleanup.completed`	DEBUG	Session cleanup completed successfully

Wildcard Patterns

Use these patterns in events filters to match groups of events.

Pattern	Description
`request.*`	All request-related events
`error.*`	All error events
`mcp.*`	All MCP/tool-related events
`agent.*`	All agent processing events
`memory.*`	All memory operation events
`response.*`	All response generation events
`a2a.*`	All agent-to-agent communication events
`performance.*`	All performance monitoring events
`clarification.*`	All clarification-related events
`planning.*`	All planning workflow events
`session.clarification.*`	All clarification session events
`tool.parameter.*`	All tool parameter collection events
`intent.*`	All intent detection events
`goal.*`	All goal tracking and achievement events
`info.*`	All information analysis and management events
`language.*`	All multilingual processing events
`quality.*`	All quality metrics and success rates
`health.*`	All system health monitoring events
`conversation.*`	All conversation flow and context events
`*`	All events (use with caution in production)

Event Selection Strategies

Use Case	Recommended Filter
Development	`["*"]` or `level: "debug"`
Production monitoring	`["request.", "response.", "error.*"]`
Performance analysis	`["performance.", "resource.", "model.inference.*"]`
Debugging issues	`["error.", ".failed", ".timeout."]`
Security auditing	`["request.denied.", "auth.", "security.*"]`
Clarification quality	`["clarification.", "planning.", "quality.*"]`
User experience	`["conversation.", "session.clarification.", "intent.*"]`
Multilingual support	`["language.", "intent.", "pattern.*"]`

Routing Configuration

Route events using logging.system for infrastructure and logging.conversation.streams for user-facing events:

logging:
  system:
    level: info
    destination: stdout

  conversation:
    enabled: true
    streams:
      - transport: stdout
        level: info
        format: jsonl
        events: ["request.*", "response.*", "error.*"]

      - transport: file
        destination: /var/log/muxi/debug.jsonl
        level: debug
        events: ["*"]

      - transport: stream
        destination: https://logs.example.com/ingest
        format: datadog_json
        events: ["error.*", "performance.*"]
        auth:
          type: bearer
          token: "${{ secrets.LOG_TOKEN }}"

See Observability for full configuration options.

We use cookies