feat: component-separated logging with session context and filtering (#7991)
* feat: component-separated logging with session context and filtering Phase 1 — Gateway log isolation: - gateway.log now only receives records from gateway.* loggers (platform adapters, session management, slash commands, delivery) - agent.log remains the catch-all (all components) - errors.log remains WARNING+ catch-all - Moved gateway.log handler creation from gateway/run.py into hermes_logging.setup_logging(mode='gateway') with _ComponentFilter Phase 2 — Session ID injection: - Added set_session_context(session_id) / clear_session_context() API using threading.local() for per-thread session tracking - _SessionFilter enriches every log record with session_tag attribute - Log format: '2026-04-11 10:23:45 INFO [session_id] logger.name: msg' - Session context set at start of run_conversation() in run_agent.py - Thread-isolated: gateway conversations on different threads don't leak Phase 3 — Component filtering in hermes logs: - Added --component flag: hermes logs --component gateway|agent|tools|cli|cron - COMPONENT_PREFIXES maps component names to logger name prefixes - Works with all existing filters (--level, --session, --since, -f) - Logger name extraction handles both old and new log formats Files changed: - hermes_logging.py: _SessionFilter, _ComponentFilter, COMPONENT_PREFIXES, set/clear_session_context(), gateway.log creation in setup_logging() - gateway/run.py: removed redundant gateway.log handler (now in hermes_logging) - run_agent.py: set_session_context() at start of run_conversation() - hermes_cli/logs.py: --component filter, logger name extraction - hermes_cli/main.py: --component argument on logs subparser Addresses community request for component-separated, filterable logging. Zero changes to existing logger names — __name__ already provides hierarchy. * fix: use LogRecord factory instead of per-handler _SessionFilter The _SessionFilter approach required attaching a filter to every handler we create. Any handler created outside our _add_rotating_handler (like the gateway stderr handler, or third-party handlers) would crash with KeyError: 'session_tag' if it used our format string. Replace with logging.setLogRecordFactory() which injects session_tag into every LogRecord at creation time — process-global, zero per-handler wiring needed. The factory is installed at import time (before setup_logging) so session_tag is available from the moment hermes_logging is imported. - Idempotent: marker attribute prevents double-wrapping on module reload - Chains with existing factory: won't break third-party record factories - Removes _SessionFilter from _add_rotating_handler and setup_verbose_logging - Adds tests: record factory injection, idempotency, arbitrary handler compat
This commit is contained in:
@ -7,16 +7,28 @@ gateway call early in their startup path. All log files live under
|
||||
Log files produced:
|
||||
agent.log — INFO+, all agent/tool/session activity (the main log)
|
||||
errors.log — WARNING+, errors and warnings only (quick triage)
|
||||
gateway.log — INFO+, gateway-only events (created when mode="gateway")
|
||||
|
||||
Both files use ``RotatingFileHandler`` with ``RedactingFormatter`` so
|
||||
All files use ``RotatingFileHandler`` with ``RedactingFormatter`` so
|
||||
secrets are never written to disk.
|
||||
|
||||
Component separation:
|
||||
gateway.log only receives records from ``gateway.*`` loggers —
|
||||
platform adapters, session management, slash commands, delivery.
|
||||
agent.log remains the catch-all (everything goes there).
|
||||
|
||||
Session context:
|
||||
Call ``set_session_context(session_id)`` at the start of a conversation
|
||||
and ``clear_session_context()`` when done. All log lines emitted on
|
||||
that thread will include ``[session_id]`` for filtering/correlation.
|
||||
"""
|
||||
|
||||
import logging
|
||||
import os
|
||||
import threading
|
||||
from logging.handlers import RotatingFileHandler
|
||||
from pathlib import Path
|
||||
from typing import Optional
|
||||
from typing import Optional, Sequence
|
||||
|
||||
from hermes_constants import get_config_path, get_hermes_home
|
||||
|
||||
@ -25,9 +37,14 @@ from hermes_constants import get_config_path, get_hermes_home
|
||||
# unless ``force=True``.
|
||||
_logging_initialized = False
|
||||
|
||||
# Default log format — includes timestamp, level, logger name, and message.
|
||||
_LOG_FORMAT = "%(asctime)s %(levelname)s %(name)s: %(message)s"
|
||||
_LOG_FORMAT_VERBOSE = "%(asctime)s - %(name)s - %(levelname)s - %(message)s"
|
||||
# Thread-local storage for per-conversation session context.
|
||||
_session_context = threading.local()
|
||||
|
||||
# Default log format — includes timestamp, level, optional session tag,
|
||||
# logger name, and message. The ``%(session_tag)s`` field is guaranteed to
|
||||
# exist on every LogRecord via _install_session_record_factory() below.
|
||||
_LOG_FORMAT = "%(asctime)s %(levelname)s%(session_tag)s %(name)s: %(message)s"
|
||||
_LOG_FORMAT_VERBOSE = "%(asctime)s - %(name)s - %(levelname)s%(session_tag)s - %(message)s"
|
||||
|
||||
# Third-party loggers that are noisy at DEBUG/INFO level.
|
||||
_NOISY_LOGGERS = (
|
||||
@ -48,6 +65,99 @@ _NOISY_LOGGERS = (
|
||||
)
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Public session context API
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
def set_session_context(session_id: str) -> None:
|
||||
"""Set the session ID for the current thread.
|
||||
|
||||
All subsequent log records on this thread will include ``[session_id]``
|
||||
in the formatted output. Call at the start of ``run_conversation()``.
|
||||
"""
|
||||
_session_context.session_id = session_id
|
||||
|
||||
|
||||
def clear_session_context() -> None:
|
||||
"""Clear the session ID for the current thread.
|
||||
|
||||
Optional — ``set_session_context()`` overwrites the previous value,
|
||||
so explicit clearing is only needed if the thread is reused for
|
||||
non-conversation work after ``run_conversation()`` returns.
|
||||
"""
|
||||
_session_context.session_id = None
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Record factory — injects session_tag into every LogRecord at creation
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
def _install_session_record_factory() -> None:
|
||||
"""Replace the global LogRecord factory with one that adds ``session_tag``.
|
||||
|
||||
Unlike a ``logging.Filter`` on a handler or logger, the record factory
|
||||
runs for EVERY record in the process — including records that propagate
|
||||
from child loggers and records handled by third-party handlers. This
|
||||
guarantees ``%(session_tag)s`` is always available in format strings,
|
||||
eliminating the KeyError that would occur if a handler used our format
|
||||
without having a ``_SessionFilter`` attached.
|
||||
|
||||
Idempotent — checks for a marker attribute to avoid double-wrapping if
|
||||
the module is reloaded.
|
||||
"""
|
||||
current_factory = logging.getLogRecordFactory()
|
||||
if getattr(current_factory, "_hermes_session_injector", False):
|
||||
return # already installed
|
||||
|
||||
def _session_record_factory(*args, **kwargs):
|
||||
record = current_factory(*args, **kwargs)
|
||||
sid = getattr(_session_context, "session_id", None)
|
||||
record.session_tag = f" [{sid}]" if sid else "" # type: ignore[attr-defined]
|
||||
return record
|
||||
|
||||
_session_record_factory._hermes_session_injector = True # type: ignore[attr-defined]
|
||||
logging.setLogRecordFactory(_session_record_factory)
|
||||
|
||||
|
||||
# Install immediately on import — session_tag is available on all records
|
||||
# from this point forward, even before setup_logging() is called.
|
||||
_install_session_record_factory()
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Filters
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
class _ComponentFilter(logging.Filter):
|
||||
"""Only pass records whose logger name starts with one of *prefixes*.
|
||||
|
||||
Used to route gateway-specific records to ``gateway.log`` while
|
||||
keeping ``agent.log`` as the catch-all.
|
||||
"""
|
||||
|
||||
def __init__(self, prefixes: Sequence[str]) -> None:
|
||||
super().__init__()
|
||||
self._prefixes = tuple(prefixes)
|
||||
|
||||
def filter(self, record: logging.LogRecord) -> bool:
|
||||
return record.name.startswith(self._prefixes)
|
||||
|
||||
|
||||
# Logger name prefixes that belong to each component.
|
||||
# Used by _ComponentFilter and exposed for ``hermes logs --component``.
|
||||
COMPONENT_PREFIXES = {
|
||||
"gateway": ("gateway",),
|
||||
"agent": ("agent", "run_agent", "model_tools", "batch_runner"),
|
||||
"tools": ("tools",),
|
||||
"cli": ("hermes_cli", "cli"),
|
||||
"cron": ("cron",),
|
||||
}
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Main setup
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
def setup_logging(
|
||||
*,
|
||||
hermes_home: Optional[Path] = None,
|
||||
@ -78,8 +188,9 @@ def setup_logging(
|
||||
Number of rotated backup files to keep.
|
||||
Defaults to 3 or the value from config.yaml ``logging.backup_count``.
|
||||
mode
|
||||
Hint for the caller context: ``"cli"``, ``"gateway"``, ``"cron"``.
|
||||
Currently used only for log format tuning (gateway includes PID).
|
||||
Caller context: ``"cli"``, ``"gateway"``, ``"cron"``.
|
||||
When ``"gateway"``, an additional ``gateway.log`` file is created
|
||||
that receives only gateway-component records.
|
||||
force
|
||||
Re-run setup even if it has already been called.
|
||||
|
||||
@ -130,6 +241,18 @@ def setup_logging(
|
||||
formatter=RedactingFormatter(_LOG_FORMAT),
|
||||
)
|
||||
|
||||
# --- gateway.log (INFO+, gateway component only) ------------------------
|
||||
if mode == "gateway":
|
||||
_add_rotating_handler(
|
||||
root,
|
||||
log_dir / "gateway.log",
|
||||
level=logging.INFO,
|
||||
max_bytes=5 * 1024 * 1024,
|
||||
backup_count=3,
|
||||
formatter=RedactingFormatter(_LOG_FORMAT),
|
||||
log_filter=_ComponentFilter(COMPONENT_PREFIXES["gateway"]),
|
||||
)
|
||||
|
||||
# Ensure root logger level is low enough for the handlers to fire.
|
||||
if root.level == logging.NOTSET or root.level > level:
|
||||
root.setLevel(level)
|
||||
@ -218,9 +341,16 @@ def _add_rotating_handler(
|
||||
max_bytes: int,
|
||||
backup_count: int,
|
||||
formatter: logging.Formatter,
|
||||
log_filter: Optional[logging.Filter] = None,
|
||||
) -> None:
|
||||
"""Add a ``RotatingFileHandler`` to *logger*, skipping if one already
|
||||
exists for the same resolved file path (idempotent).
|
||||
|
||||
Parameters
|
||||
----------
|
||||
log_filter
|
||||
Optional filter to attach to the handler (e.g. ``_ComponentFilter``
|
||||
for gateway.log).
|
||||
"""
|
||||
resolved = path.resolve()
|
||||
for existing in logger.handlers:
|
||||
@ -236,6 +366,8 @@ def _add_rotating_handler(
|
||||
)
|
||||
handler.setLevel(level)
|
||||
handler.setFormatter(formatter)
|
||||
if log_filter is not None:
|
||||
handler.addFilter(log_filter)
|
||||
logger.addHandler(handler)
|
||||
|
||||
|
||||
|
||||
Reference in New Issue
Block a user