Diagnostics Flags
Diagnostics flags let you enable targeted debug logs without turning on verbose logging everywhere. Flags are opt-in and have no effect unless Fased code checks that exact flag. Use flags as an operator diagnostic tool:- Enable the narrow flag in Advanced > Config or with the environment override below.
- Reproduce the issue.
- Read the output in Logs or
fased logs --follow.
How it works
- Flags are strings (case-insensitive).
- You can enable flags in config or via an env override.
- Config flags and env flags are merged.
- Wildcards are supported:
telegram.*matchestelegram.http*enables all flags
| Flag | Emits |
|---|---|
telegram.http | Telegram send failures with HTTP status, Telegram error code, retry-after, and description. |
Enable via Advanced Config
Env override (one-off)
1, true, all, or * to enable every code-backed diagnostic flag for
that process.
Disable env-provided flags for that process:
false, off, and none are also treated as disabled. This does not erase
flags already stored in diagnostics.flags. Remove stored config flags from
Advanced > Config if you want no configured flags.
Where Logs Show Up
Flags emit logs into the standard gateway log file. By default this is the current dated file in the Fased temp log directory, normally:logging.file, use that path instead. Logs are JSONL (one JSON
object per line). Redaction still applies based on logging.redactSensitive.
The browser Logs page tails the same file through the Gateway. Use
Advanced > Debug for status snapshots and raw RPC checks; use Logs for
streaming log evidence.
Extract logs
Filter for Telegram HTTP diagnostics:fased logs --follow (see
/cli/logs).
Notes
- Telegram HTTP diagnostic logs are emitted at
warn. Iflogging.leveliserror,fatal, orsilent, they may be suppressed. Defaultinfois fine. - Flags are usually safe to leave enabled, but they can increase log volume for the specific subsystem.
- Structured diagnostic events are separate. Set
diagnostics.enabled: truefor event snapshots in Advanced > Debug and for diagnostic event subscribers. - Cache trace is separate. Use
diagnostics.cacheTrace.enabledorFASED_CACHE_TRACE=1only while debugging cache/context behavior because it can write prompt, message, and system text to JSONL. - OpenTelemetry export is separate. It requires the
diagnostics-otelextension plusdiagnostics.enabledanddiagnostics.otel.enabled. - Use Gateway logging to change log destinations, levels, and redaction.