Get desktop application:
View/edit binary Protocol Buffers messages
Emitted immediately at binary launch, before onboarding or thread creation.
Anonymous UUID from ~/.dbt/.user.yml (created on first run).
dbt Cloud account ID (empty if unauthenticated).
Whether this is the first time the binary has run (no .user.yml existed).
How the app was launched: "cli", "vscode", "desktop", "mcp_server".
Unix timestamp of launch.
Whether the WIZARD_INTERNAL env var was set for this session.
Emitted on structured errors (sidecar crashes, provider auth failures, panics).
Empty if error occurs before session starts.
"sidecar_startup" | "provider_auth" | "thread_start" | "turn_failure" | "panic" | "connection_lost"
Structured code: "health_timeout", "401", "connection_refused", etc.
Rust error type name (no message content).
"litellm_sidecar" | "app_server" | "tui" | "core" | "provider" | "cloud_requirements" | "mcp"
"anthropic" | "openai" | "bedrock" | "" if not provider-related.
Whether the WIZARD_INTERNAL env var was set for this session.
Stable anonymous client id, present even before a session exists.
Emitted when a user submits feedback via /feedback.
Unique identifier for this event.
Session correlation ID (thread_id).
Feedback category: "bad_result", "good_result", "bug", "safety_check", "other".
Optional user-provided note describing the feedback.
Deprecated; feedback no longer uploads logs.
Authenticated dbt Cloud user ID (JWT sub claim); falls back to anonymous UUID from ~/.dbt/.user.yml when unauthenticated.
Numeric dbt Cloud account ID (from JWT claims or dbt_cloud.yml).
Global dbt Cloud account identifier slug (from /api/v2/accounts/{id}/).
dbt-wizard binary version.
Operating system: "macos", "linux", "windows".
CPU architecture: "aarch64", "x86_64".
Turn correlation ID (which turn the feedback is about).
Model slug active during the session (e.g. "claude-sonnet-4-6").
Whether the WIZARD_INTERNAL env var was set for this session.
Emitted once per onboarding step transition.
Step name: "auth", "trust_directory", "dbt_project_setup", "provider_setup".
Step outcome: "started", "completed", "skipped", "abandoned".
Time spent on this step.
First time user ever onboards (no prior .user.yml).
Auth step: "dbt_platform" | "api_key" | "chatgpt"
Provider step: "anthropic" | "openai" | "bedrock" | "azure" | "snowflake" | "dbt"
Trust step: "trust" | "deny"
dbt project step: whether a dbt_project.yml was found
Position in the flow (0-based)
Total steps shown to this user
Whether the WIZARD_INTERNAL env var was set for this session.
Emitted once at session start and once at session end.
Unique identifier for this event.
Session correlation ID (thread_id).
"start" or "end".
Client name: "cli", "vscode", "desktop", "mcp_server".
Model slug (e.g. "claude-sonnet-4-20250514").
dbt-wizard binary version.
Operating system: "macos", "linux", "windows".
CPU architecture: "aarch64", "x86_64".
Authenticated dbt Cloud user ID (JWT sub claim); falls back to anonymous UUID from ~/.dbt/.user.yml when unauthenticated.
0 on start; total session wall-clock time on end.
Empty on start; "completed", "failed", or "interrupted" on end.
0 on start; total turn count on end.
0 on start; cumulative token usage on end.
Numeric dbt Cloud account ID (from JWT claims or dbt_cloud.yml).
Global dbt Cloud account identifier slug (from /api/v2/accounts/{id}/).
Whether the WIZARD_INTERNAL env var was set for this session.
Emitted at subagent start and end to track full lifecycle.
Parent session (the main thread_id).
The subagent's own thread_id.
"review" | "compact" | "thread_spawn" | "memory_consolidation" | "other"
"start" | "end"
Only on "end": "completed" | "failed" | "interrupted"
0 on start; populated on end.
Nesting depth (0 = direct child of main session).
Whether the WIZARD_INTERNAL env var was set for this session.
Emitted once per tool invocation.
Unique identifier for this event.
Session correlation ID (thread_id).
Turn correlation ID.
Tool category: "builtin", "mcp", "shell", "subagent".
Specific tool name (e.g. "dbt_index.search", "apply_patch").
Whether the tool call resulted in an error.
Tool execution wall-clock time.
Whether the WIZARD_INTERNAL env var was set for this session.
Emitted once per turn completion.
Unique identifier for this event.
Session correlation ID (thread_id).
Turn correlation ID.
Model slug used for this turn.
Turn outcome: "completed", "failed", "interrupted".
Turn wall-clock duration.
Token usage breakdown.
Number of tool calls made during this turn.
Whether this is the first turn in the session.
Collaboration mode: "Default", "Plan", etc.
Whether the WIZARD_INTERNAL env var was set for this session.