Logging

Logging Levels

The Flopsar server uses logging levels. The levels are defined as follows:

debug : more detailed logging than info. Used to log additional information about operation and data processing. Use only when you suspect a problem, want to trace request handling, or Flopsar Technology Support asks you to. It prints a large amount of diagnostic information.

info : default level. Used to log runtime information such as configuration checks. Use this level in production.

warn : an error occurred with no impact on stability or data integrity.

error : an error that can affect operation or data integrity. System-call and library errors are also logged at this level.

Messages

Logger messages have the following format:

DATETIME LEVEL [XXXXXXXX] LOG_ENTRY

where:

DATETIME - date and time of the log entry,

LEVEL - log level,

XXXXXXXX - first eight characters of the agent instance name,

LOG_ENTRY - log entry

Server

If the logger backend is Local, server logs are stored under <home_directory>/logs. If the backend is Syslog, read them with:

$ journalctl -t flopsar -r

On a packaged install, journalctl -u flopsar is usually the first place to look. Logger backend, file size, and rotation are set in [logger] — see the Server Configuration Reference.

Security-relevant events are a separate, tamper-evident trail. See Logging, Monitoring & Audit Trail.

If Log model requests is on under Administration → AI Assistant, the server also writes provider request bodies to <home_directory>/logs/llm.log (rotated at 50 MB to llm.log.1). That file is opt-in, is not the audit trail, and contains prompts and tool results. Leave it off unless you are debugging a provider integration. See AI Assistant.

The agent does not write log files on disk. Internal agent errors appear on the instance Diagnostics tab in the Workstation — see Agents.