flue logs

Synopsis

flue logs <workflowRunId> [--server <url>] [--follow|-f|--no-follow] [--since <eventIndex>] [--types <a,b,c>] [--limit <n>] [--format <pretty|json|ndjson>]

Description

flue logs replays or follows events for one workflow run from a running Flue server. It is read-only and does not invoke work.

Runs are workflow-only. Direct HTTP or WebSocket agent prompts and dispatched agent inputs are persistent session interactions, not runs.

flue logs inspects runs owned by the selected running server. It cannot inspect the private child process used by flue run after that command exits.

Arguments

Argument	Description
<workflowRunId>	Workflow run ID to inspect.

Options

Option	Default	Description
--server <url>	http://127.0.0.1:3583	Select the running Flue server base URL beneath which /runs/<runId> is published. May include an authored mount prefix.
--follow, -f	Automatic	Force live event streaming.
--no-follow	Automatic	Replay persisted events once and exit.
--since <eventIndex>	Beginning of history	Resume strictly after an event index.
--types <a,b,c>	All event types	Emit only the selected comma-separated event types.
--limit <n>	Replay: 100; follow: unlimited	Limit emitted events. Replay requests accept at most 1000.
--format <pretty|json|ndjson>	pretty	Select human-readable or line-delimited JSON output.

When neither follow option is passed, flue logs follows active runs and requests one persisted-event snapshot for terminal runs. The snapshot endpoint is bounded: it returns the first 100 matching events by default and accepts a maximum --limit of 1000. flue logs does not paginate snapshots.

Run routes may expose sensitive payloads, results, errors, and events. The current CLI does not accept authentication headers; use it only with a suitably protected and reachable inspection endpoint.

Output and exit behavior

pretty writes human-readable events to stderr. json and ndjson are currently aliases: each writes one JSON event object per stdout line.

Request failures exit with status 1. A failed workflow exits with status 2 only when its failing run_end event is consumed. Stream-framed server failures are not currently guaranteed to produce a nonzero exit status.

Examples

flue logs run_01H...
flue logs run_01H... --no-follow
flue logs run_01H... --since 25
flue logs run_01H... --types operation_start,operation,tool_call,log,run_end --format ndjson

See Observability for workflow-run inspection and telemetry guidance.