Usage ===== Once ``django_logging`` is installed and added to your ``INSTALLED_APPS``, you can start using it right away. The package provides several features to customize and enhance logging in your Django project. Below is a guide on how to use the various features provided by `django_logging`. Basic Logging Usage ------------------- At its core, `django_logging` is built on top of Python’s built-in logging module. This means you can use the standard logging module to log messages across your Django project. Here’s a basic example of logging usage: .. code-block:: python import logging logger = logging.getLogger(__name__) logger.debug("This is a debug message") logger.info("This is an info message") logger.warning("This is a warning message") logger.error("This is an error message") logger.critical("This is a critical message") These logs will be handled according to the configurations set up by `django_logging`, using either the default settings or any custom settings you've provided. Context Manager --------------- You can use the ``config_setup`` context manager to temporarily apply `django_logging` configurations within a specific block of code. Example usage: .. code-block:: python from django_logging.utils.context_manager import config_setup import logging logger = logging.getLogger(__name__) def foo(): logger.info("This log will use the configuration set in the context manager!") with config_setup(): """Your logging configuration changes here""" foo() Note: ``AUTO_INITIALIZATION_ENABLE`` must be set to ``False`` in the settings to use the context manager. If it is ``True``, attempting to use the context manager will raise a ``ValueError`` with the message: .. code-block:: text "You must set 'AUTO_INITIALIZATION_ENABLE' to False in DJANGO_LOGGING in your settings to use the context manager." Log and Notify Utility ---------------------- To send specific logs as email, use the ``log_and_notify_admin`` function. Ensure that the ``ENABLE`` option in ``LOG_EMAIL_NOTIFIER`` is set to ``True`` in your settings: .. code-block:: python from django_logging.utils.log_email_notifier.log_and_notify import log_and_notify_admin import logging logger = logging.getLogger(__name__) log_and_notify_admin(logger, logging.INFO, "This is a log message") You can also include additional request information (``ip_address`` and ``browser_type``) in the email by passing an ``extra`` dictionary: .. code-block:: python from django_logging.utils.log_email_notifier.log_and_notify import log_and_notify_admin import logging logger = logging.getLogger(__name__) def some_view(request): log_and_notify_admin( logger, logging.INFO, "This is a log message", extra={"request": request} ) Note: To use the email notifier, ``LOG_EMAIL_NOTIFIER["ENABLE"]`` must be set to ``True``. If it is not, calling ``log_and_notify_admin`` will raise a ``ValueError``: .. code-block:: text "Email notifier is disabled. Please set the 'ENABLE' option to True in the 'LOG_EMAIL_NOTIFIER' in DJANGO_LOGGING in your settings to activate email notifications." Additionally, ensure that all required email settings are configured in your Django settings file. - **Note**: For more detailed configuration options, refer to the :doc:`Settings `. Execution Tracker ----------------- The ``execution_tracker`` decorator is used to log the performance metrics of any function. It tracks execution time and the number of database queries for decorated function (if enabled). Example Usage: .. code-block:: python from django_logging.decorators import execution_tracker @execution_tracker( logging_level=logging.INFO, log_queries=True, query_threshold=10, query_exceed_warning=False, ) def some_function(): # function code pass **Arguments:** - ``logging_level`` (``int``): The logging level at which performance details will be logged. Defaults to ``logging.INFO``. - ``log_queries`` (``bool``): Whether to log the number of database queries for decorated function(if ``DEBUG`` is ``True`` in your settings). If ``log_queries=True``, the number of queries will be included in the logs. Defaults to ``False``. - ``query_threshold`` (``int``): If provided, the number of database queries will be checked. If the number of queries exceeds the given threshold, a warning will be logged. Defaults to ``None``. - ``query_exceed_warning`` (``bool``): Whether to log a ``WARNING`` message if the number of queries exceeds the threshold. Defaults to ``False``. **Example Log Output:** .. code-block:: shell INFO | 'datetime' | execution_tracking | Performance Metrics for Function: 'some_function' Module: some_module File: /path/to/file.py, Line: 123 Execution Time: 0.21 second(s) Database Queries: 15 queries (exceeds threshold of 10) If `log_queries` is set to ``True`` but ``DEBUG`` is ``False``, a ``WARNING`` will be logged: .. code-block:: shell WARNING | 'datetime' | execution_tracking | DEBUG mode is disabled, so database queries are not tracked. To include the number of queries, set ``DEBUG`` to ``True`` in your Django settings. Request Logging Middleware -------------------------- The ``django_logging.middleware.RequestLogMiddleware`` is a middleware that logs detailed information about each incoming request to the server. It is capable of handling both synchronous and asynchronous requests. To enable this middleware, add it to your Django project's ``MIDDLEWARE`` setting: .. code-block:: python MIDDLEWARE = [ # ... "django_logging.middleware.RequestLogMiddleware", # ... ] Key Features ^^^^^^^^^^^^ 1. **Request Information Logging**: - Logs the following details at the start of each request: - Request method - Request path - Query parameters - Referrer (if available) - Example log at request start: .. code-block:: text INFO | 2024-10-03 16:29:47 | request_middleware | REQUEST STARTED: method=GET path=/admin/ query_params=None referrer=http://127.0.0.1:8000/admin/login/?next=/admin/ | {'ip_address': '127.0.0.1', 'request_id': '09580021-6bff-4b82-99b5-c52406b2cc91', 'user_agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/129.0.0.0 Safari/537.36'} 2. **Response Information Logging**: - Logs the following details after the request is processed: - User (or 'Anonymous' if not authenticated) - HTTP Status code - Content type - Time taken to process the request - Optionally logs SQL queries executed during the request (if enabled) - Example log at request completion: .. code-block:: text INFO | 2024-10-03 16:29:47 | request_middleware | REQUEST FINISHED: user=[mehrshad (ID:1)] status_code=200 content_type=[text/html; charset=utf-8] response_time=[0.08 second(s)] 3 SQL QUERIES EXECUTED Query1={'Time': 0.000(s), 'Query': [SELECT "django_session"."session_key", "django_session"."session_data", "django_session"."expire_date" FROM "django_session" WHERE ("django_session"."expire_date" > '2024-10-03 12:59:47.812918' AND "django_session"."session_key" = 'uq0nrbglazfm4cy656w3451xydfirh45') LIMIT 21]} Query2={'Time': 0.001(s), 'Query': [SELECT "auth_user"."id", "auth_user"."password", "auth_user"."last_login", "auth_user"."is_superuser", "auth_user". "username", "auth_user"."first_name", "auth_user"."last_name", "auth_user"."email", "auth_user"."is_staff", "auth_user". "is_active", "auth_user"."date_joined" FROM "auth_user" WHERE "auth_user"."id" = 1 LIMIT 21]} | {'ip_address': '127.0.0.1', 'request_id': '09580021-6bff-4b82-99b5-c52406b2cc91', 'user_agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/129.0.0.0 Safari/537.36'} 3. **Request ID**: - A unique request ID is generated for each request (or taken from the ``X-Request-ID`` header if provided). This request ID is included in logs and can help with tracing specific requests. 4. **SQL Query Logging**: - If SQL query logging is enabled, all queries executed during the request will be logged along with their execution time. **Note**: to enable Query logging, you can set ``LOG_SQL_QUERIES_ENABLE`` to ``True`` in settings. for more details, refer to the :doc:`Settings `. 5. **Streaming Response Support**: - The middleware supports both synchronous and asynchronous streaming responses, logging the start and end of the streaming process, as well as any errors during streaming. 6. **User Information**: - Logs the authenticated user's username and ID if available, or 'Anonymous' if the user is not authenticated. 7. **IP Address and User-Agent**: - The middleware logs the client's IP address and user agent for each request. Context Variables Usage ^^^^^^^^^^^^^^^^^^^^^^^^ We use context variables in ``RequestLogMiddleware`` to store the following information for each request: - **request_id**: A unique identifier for the request. - **ip_address**: The client’s IP address. - **user_agent**: The client's user agent string. These context variables can be accessed and used in other parts of the logging system or during the request processing lifecycle. MonitorLogSizeMiddleware ------------------------ This middleware monitors the size of the log directory and checks it weekly. It triggers the ``logs_size_audit`` management command to assess the total size of the log files. If the log directory size exceeds a certain limit (``LOG_DIR_SIZE_LIMIT``), the middleware sends a warning email to the ``ADMIN_EMAIL`` asynchronously. To enable this middleware, add it to your Django project's ``MIDDLEWARE`` setting: .. code-block:: python MIDDLEWARE = [ # ... "django_logging.middleware.MonitorLogSizeMiddleware", # ... ] Context Variable Management --------------------------- ``django_logging`` includes a powerful ``ContextVarManager`` class, allowing you to manage context variables dynamically within your logging system. These variables are bound to the current context and automatically included in your log entries via the ``%(context)s`` placeholder in the log format. Binding and Unbinding Context Variables ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ The ``ContextVarManager`` provides several methods to manage context variables efficiently: ``bind(**kwargs)`` ~~~~~~~~~~~~~~~~~~ The ``bind`` method allows you to bind key-value pairs as context variables that will be available during the current context. These variables can be used to add contextual information to log entries. **Example**: .. code-block:: python from django_logging.contextvar import manager import logging logger = logging.getLogger(__name__) # Binding context variables manager.bind(user="test_user", request_id="abc123") logger.info("Logging with context") **Log Output**: .. code-block:: text INFO | 2024-10-03 12:00:00 | Logging with context | {'user': 'test_user', 'request_id': 'abc123'} ``unbind(key: str)`` ~~~~~~~~~~~~~~~~~~~~ The ``unbind`` method removes a specific context variable by its key. It effectively clears the context variable from the log entry. **Example**: .. code-block:: python manager.unbind("user") logger.info("Logging without the 'user' context") **Log Output**: .. code-block:: text INFO | 2024-10-03 12:05:00 | Logging without the 'user' context | {'request_id': 'abc123'} Batch Binding and Resetting Context Variables ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ``batch_bind(**kwargs)`` ~~~~~~~~~~~~~~~~~~~~~~~~ The ``batch_bind`` method binds multiple context variables at once and returns tokens that can be used later to reset the variables to their previous state. This is useful when you need to bind a group of variables temporarily. **Example**: .. code-block:: python tokens = manager.batch_bind(user="admin_user", session_id="xyz789") logger.info("Logging with batch-bound context") **Log Output**: .. code-block:: text INFO | 2024-10-03 12:10:00 | Logging with batch-bound context | {'user': 'admin_user', 'session_id': 'xyz789'} ``reset(tokens: Dict[str, contextvars.Token])`` ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The ``reset`` method allows you to reset context variables to their previous state using tokens returned by ``batch_bind``. **Example**: .. code-block:: python manager.reset(tokens) logger.info("Context variables have been reset") **Log Output**: .. code-block:: text INFO | 2024-10-03 12:15:00 | Context variables have been reset | ``clear()`` ~~~~~~~~~~~ The ``clear`` method clears all bound context variables at once, effectively removing all contextual data from the log entry. **Example**: .. code-block:: python manager.clear() logger.info("All context variables cleared") **Log Output**: .. code-block:: text INFO | 2024-10-03 12:20:00 | All context variables cleared | Retrieving and Merging Context Variables ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ``get_contextvars()`` ~~~~~~~~~~~~~~~~~~~~~ The ``get_contextvars`` method retrieves the current context variables available in the context. This method is useful for inspecting or debugging the context. **Example**: .. code-block:: python current_context = manager.get_contextvars() print(current_context) # Output: {'user': 'admin_user', 'session_id': 'xyz789'} ``merge_contexts(bound_context: Dict[str, Any], local_context: Dict[str, Any])`` ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The ``merge_contexts`` method merges two dictionaries of context variables, giving priority to the ``bound_context``. This is useful when you want to combine different sources of context data. **Example**: .. code-block:: python bound_context = {"user": "admin_user"} local_context = {"request_id": "12345"} merged_context = manager.merge_contexts(bound_context, local_context) print(merged_context) # Output: {'user': 'admin_user', 'request_id': '12345'} ``get_merged_context(bound_logger_context: Dict[str, Any])`` ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The ``get_merged_context`` method combines the bound logger context with the current context variables, allowing you to retrieve a single dictionary with all the context data. **Example**: .. code-block:: python bound_logger_context = {"app_name": "my_django_app"} merged_context = manager.get_merged_context(bound_logger_context) print(merged_context) # Output: {'app_name': 'my_django_app', 'user': 'admin_user'} Scoped Context Management ^^^^^^^^^^^^^^^^^^^^^^^^^ ``scoped_context(**kwargs)`` ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The ``scoped_context`` method provides a context manager to bind context variables temporarily for a specific block of code. After the block completes, the context variables are automatically unbound. **Example**: .. code-block:: python with manager.scoped_context(transaction_id="txn123"): logger.info("Scoped context active") logger.info("Scoped context no longer active") **Log Output**: .. code-block:: text INFO | 2024-10-03 12:30:00 | Scoped context active | {'transaction_id': 'txn123'} INFO | 2024-10-03 12:30:10 | Scoped context no longer active | send_logs Command ----------------- This Django management command zips the log directory and emails it to a specified email address. The command is useful for retrieving logs remotely and securely, allowing administrators to receive log files via email. Key Features ^^^^^^^^^^^^ - **Zips Log Directory**: Automatically compresses the log directory into a single zip file. - **Email Log Files**: Sends the zipped log file to a specified email address. How It Works ^^^^^^^^^^^^ 1. **Setup Log Directory**: The command retrieves the log directory from Django settings (``DJANGO_LOGGING['LOG_DIR']``). 2. **Zip the Logs**: Compresses the entire log directory into a zip file stored in a temporary location. 3. **Email the Zip File**: Sends the zipped log file to the email address provided as an argument, attaching it to an email message. 4. **Cleanup**: After sending the email, the temporary zip file is deleted to free up disk space. Command Execution ^^^^^^^^^^^^^^^^^ To execute the command, use the following syntax: .. code-block:: shell python manage.py send_logs Example ^^^^^^^ If you want to send the logs to `admin@example.com`, the command would be: .. code-block:: shell python manage.py send_logs admin@example.com This will zip the log directory and send it to `admin@example.com` with the subject "Log Files". generate_pretty_json Command ---------------------------- This Django management command allows you to locate and prettify JSON log files stored in a log directory that generated by ``django_logging``. It takes JSON files from the log directory, formats them into a clean, readable structure, and saves the result in the ``pretty`` directory. Key Features ^^^^^^^^^^^^ - **Locate JSON Logs**: Automatically finds ``.json`` files in the `json` log directory. - **Pretty Formatting**: Reformats the JSON logs into a valid JSON array with proper indentation, improving readability. - **Separate Output Directory**: Saves the reformatted JSON files in a ``pretty`` subdirectory, preserving the original files. How It Works ^^^^^^^^^^^^ 1. **Setup Directories**: The command looks for a ``json`` subdirectory within your defined log directory. If it doesn't exist, an error is displayed. 2. **Process JSON Files**: Each `.json` file found in the directory is processed: - Parses multiple JSON objects within the file. - Reformats them as a pretty JSON array with proper indentation. - Saves the new, formatted JSON in the ``pretty`` subdirectory with the prefix ``formatted_``. Command Execution ^^^^^^^^^^^^^^^^^ To execute the command, use the following syntax: .. code-block:: bash python manage.py generate_pretty_json Example ^^^^^^^ Running the command will process the following files: - ``logs/json/error.json`` ➡ ``logs/json/pretty/formatted_error.json`` - ``logs/json/info.json`` ➡ ``logs/json/pretty/formatted_info.json`` generate_pretty_xml Command --------------------------- This Django management command allows you to locate and reformat XML log files stored in a log directory generated by ``django_logging``. It processes XML files by wrapping their content in a `` element and saves the reformatted files in a separate directory. Key Features ^^^^^^^^^^^^ - **Locate XML Logs**: Automatically finds ``.xml`` files in the ``xml`` log directory. - **Reformatting**: Wraps XML content in a `` element, ensuring consistency in structure. - **Separate Output Directory**: Saves the reformatted XML files in a ``pretty`` subdirectory with the prefix ``formatted_``, preserving the original files. How It Works ^^^^^^^^^^^^ 1. **Setup Directories**: The command looks for an ``xml`` subdirectory within your defined log directory. If it doesn't exist, an error is displayed. 2. **Process XML Files**: Each `.xml` file found in the directory is processed: - The content of each XML file is wrapped inside a `` element. - The reformatted file is saved in the ``pretty`` subdirectory with the prefix ``formatted_``. Command Execution ^^^^^^^^^^^^^^^^^ To execute the command, use the following syntax: .. code-block:: bash python manage.py generate_pretty_xml Example ^^^^^^^ Running the command will process the following files: - ``logs/xml/error.xml`` ➡ ``logs/xml/pretty/formatted_error.xml`` - ``logs/xml/info.xml`` ➡ ``logs/xml/pretty/formatted_info.xml`` logs_size_audit Command -------------------------- This Django management command monitors the size of your log directory. If the total size exceeds the configured limit, the command sends a warning email notification to the admin. It also prints a per-category breakdown showing the size of active, rotated, compressed, and archived files separately. Key Features ^^^^^^^^^^^^ - **Log Directory Size Check**: Calculates the total size of the log directory. - **Per-Category Breakdown**: Separates file sizes into four buckets: active log files, rotated backups, gzip-compressed files, and archived bundles. - **Configurable Size Limit**: Compares the total size against ``LOG_DIR_SIZE_LIMIT`` in settings. - **Email Notification**: Sends a warning email to the administrator when the limit is exceeded, including the full breakdown in the email body. How It Works ^^^^^^^^^^^^ 1. **Setup Log Directory**: Retrieves the log directory from ``DJANGO_LOGGING['LOG_DIR']``. If the directory doesn't exist, an error is displayed. 2. **Categorize Files**: Classifies each file as active (``.log``), rotated (``.log.N``, etc.), compressed (``.gz``), or archived (inside ``archive/``). 3. **Compare with Size Limit**: Compares the total size in MB with ``LOG_DIR_SIZE_LIMIT``. 4. **Send Warning Email**: If the limit is exceeded, sends a warning email to ``ADMIN_EMAIL`` that includes the full breakdown. Command Execution ^^^^^^^^^^^^^^^^^ To execute the command, use the following syntax: .. code-block:: shell python manage.py logs_size_audit Example ^^^^^^^ Running the command produces a size breakdown: .. code-block:: text Log directory size breakdown: active 5 file(s) 0.42 MB rotated 12 file(s) 18.30 MB compressed 8 file(s) 4.10 MB archived 3 file(s) 2.00 MB TOTAL 24.82 MB If the total exceeds ``LOG_DIR_SIZE_LIMIT``, an email is sent to ``ADMIN_EMAIL`` configured in Django settings. rotate_logs Command ------------------- This management command manually triggers an immediate rollover on all active rotating log handlers. It is useful in deployment scripts to start fresh log files right after a release, while safely preserving old log content in numbered backup files. Key Features ^^^^^^^^^^^^ - **Immediate rollover**: calls ``doRollover()`` on every live ``RotatingFileHandler`` or ``TimedRotatingFileHandler`` found across all active loggers. - **Graceful skip**: plain ``FileHandler`` instances (when ``LOG_ROTATION TYPE`` is ``"none"``) are reported as skipped rather than causing an error. - **Multi-handler support**: rotates all qualifying handlers in a single pass, regardless of which logger they are attached to. How It Works ^^^^^^^^^^^^ 1. Iterates over the root logger and all named loggers. 2. For each handler that is a ``RotatingFileHandler`` or ``TimedRotatingFileHandler``, acquires the handler lock, calls ``doRollover()``, and releases the lock. 3. Reports rotated file paths and skipped plain handlers to stdout. Command Execution ^^^^^^^^^^^^^^^^^ .. code-block:: shell python manage.py rotate_logs Example Output ^^^^^^^^^^^^^^ When rotation succeeds: .. code-block:: text Rotated 5 handler(s): /var/log/myapp/debug.log /var/log/myapp/info.log ... When no rotating handlers are configured: .. code-block:: text No rotating handlers found. Set LOG_ROTATION TYPE to 'size' or 'time' to enable rotation. archive_logs Command -------------------- This management command moves rotated log files into a timestamped subdirectory under ``logs/archive/``. Active ``.log`` files are never touched. Optionally, individual files can be gzip-compressed before moving, and the entire archive directory can be bundled into a single ``tar.gz`` or ``zip`` file. Key Features ^^^^^^^^^^^^ - **Safe archiving**: only rotated files (``debug.log.1``, ``info.log.2024-01-15``, etc.) are moved. Active ``.log`` files remain in place. - **``--compress``**: gzip-compresses each uncompressed rotated file before archiving (``info.log.1`` → ``info.log.1.gz``). - **``--bundle``**: after archiving, bundles the entire timestamped directory into a single ``tar.gz`` or ``zip`` file and removes the directory. - **``--dry-run``**: prints exactly what would happen without touching any files. - **Re-run safe**: existing ``archive/`` subdirectories are never re-archived on subsequent runs. How It Works ^^^^^^^^^^^^ 1. Scans the log directory for files that do not match the active ``.log`` naming pattern. 2. Creates ``/archive//`` and moves the rotated files into it. 3. If ``--compress`` is set, uncompressed files are gzip-compressed during the move. 4. If ``--bundle`` is set, the timestamped directory is compressed into a single file and then deleted. Command Execution ^^^^^^^^^^^^^^^^^ .. code-block:: shell # Basic — move rotated files into archive/ python manage.py archive_logs # Gzip each file before moving python manage.py archive_logs --compress # Bundle archive directory into a tar.gz and remove the directory python manage.py archive_logs --bundle tar.gz # Bundle into a zip file python manage.py archive_logs --bundle zip # Compress individual files and then bundle into tar.gz python manage.py archive_logs --compress --bundle tar.gz # Preview everything without touching any files python manage.py archive_logs --compress --bundle tar.gz --dry-run Arguments ^^^^^^^^^ ``--compress`` Gzip-compress each uncompressed rotated file before moving it to the archive directory. Files that are already ``.gz`` are moved as-is. ``--bundle `` After archiving, bundle the entire timestamped archive directory into a single compressed file and remove the source directory. Supported formats: ``tar.gz``, ``zip``. ``--dry-run`` Print what would be archived and bundled without moving or creating any files. Example Output ^^^^^^^^^^^^^^ .. code-block:: text Archived 3 file(s) to /var/log/myapp/archive/2024-03-15_143022: /var/log/myapp/archive/2024-03-15_143022/debug.log.1.gz /var/log/myapp/archive/2024-03-15_143022/info.log.2024-03-14.gz /var/log/myapp/archive/2024-03-15_143022/error.log.1.gz Bundled archive directory into /var/log/myapp/archive/2024-03-15_143022.tar.gz (0.04 MB) and removed the source directory. Resulting layout after ``--bundle tar.gz``: .. code-block:: text logs/ debug.log ← active, untouched info.log ← active, untouched archive/ 2024-03-15_143022.tar.gz ← all rotated files, bundled Log Rotation ------------ ``django_logging`` supports dynamic log file rotation through the ``LOG_ROTATION`` setting. By default no rotation is applied. Setting ``TYPE`` to ``"size"`` or ``"time"`` switches the underlying handler automatically — no other code changes are needed. For full configuration details refer to the :doc:`Settings ` documentation. Size-based rotation example ^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. code-block:: python DJANGO_LOGGING = { "LOG_ROTATION": { "TYPE": "size", "MAX_BYTES": 5_000_000, # 5 MB "BACKUP_COUNT": 10, "COMPRESS": True, # produces debug.log.1.gz, info.log.1.gz … }, } Time-based rotation example ^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. code-block:: python DJANGO_LOGGING = { "LOG_ROTATION": { "TYPE": "time", "WHEN": "midnight", "INTERVAL": 1, "BACKUP_COUNT": 30, }, } Per-level rotation overrides ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Use ``LOG_ROTATION_OVERRIDES`` to give individual log levels a different strategy. Only the keys you specify are changed; everything else inherits from ``LOG_ROTATION``. .. code-block:: python DJANGO_LOGGING = { "LOG_ROTATION": { "TYPE": "size", "MAX_BYTES": 5_000_000, "BACKUP_COUNT": 5, }, "LOG_ROTATION_OVERRIDES": { "ERROR": {"TYPE": "time", "WHEN": "midnight", "BACKUP_COUNT": 30}, "CRITICAL": {"TYPE": "time", "WHEN": "midnight", "BACKUP_COUNT": 90}, }, }