Skip to content

cmem_client.components.explore.logs¤

The in-memory log buffer of Explore (DataPlatform).

Classes:

  • ExploreLogs – Read the in-memory log buffer of Explore: the most recent lines, and its state.

ExploreLogs¤

Bases: _LogBuffer

Read the in-memory log buffer of Explore: the most recent lines, and its state.

Each Explore instance buffers its own lines, so a response describes exactly one instance, and its sequence numbers are only meaningful together with the instance identifier it reports.

Examples:

>>> from cmem_client.client import Client
>>> client = Client.from_env()
>>> tail = client.explore.logs.get_tail(level="WARN", limit=10)
>>> [line.message for line in tail.lines]

Functions:

  • get_status – Read the state of the in-memory log buffer of Explore.
  • get_tail – Read the most recent log lines of Explore.

Attributes:

get_status¤

get_status()

Read the state of the in-memory log buffer of Explore.

Reports whether lines are captured at all, at which level, how many are held and which loggers are excluded. This is the way to find out why a log read returns fewer lines than expected, or none.

Returns:

  • LogBufferStatus – The state and configuration of the buffer of the instance which answered.

Raises:

  • HTTPStatusError – If the request fails, for example because the deployment does not know the endpoint or the user is not authorized for it.
  • LogsReadError – If the answer is not a log buffer status.

get_tail¤

get_tail(since=None, limit=None, level=None, logger=None, contains=None, request_id=None)

Read the most recent log lines of Explore.

The lines are returned oldest first, and every filter is applied by Explore before the limit. Which end the limit cuts depends on since: without it Explore answers the newest limit matching lines, and with it the oldest limit lines after the cursor, so a follower pages forward instead of skipping ahead.

To follow the log, pass the last_sequence of the previous response as since of the next request. A dropped above zero in a response means the client fell behind and those lines are no longer buffered. Sequence numbers belong to the instance which answered, so a changed instance_id invalidates the cursor.

Parameters:

  • since (int | None) – Return only lines with a sequence higher than this one.
  • limit (int | None) – Maximum number of lines to return. Explore clamps this to its own configured maximum rather than rejecting a larger value, so asking for an arbitrarily high number is how to read as much as it will give. Without a limit Explore answers its own default page, which is smaller than its maximum.
  • level (LogLevel | str | None) – Minimum level a line has to be logged at. Asking for WARN returns WARN and ERROR lines.
  • logger (list[str] | None) – Return only lines of loggers starting with one of these prefixes.
  • contains (list[str] | None) – Return only lines containing all of these substrings, ignoring case.
  • request_id (str | None) – Return only lines produced while handling the request with this correlation identifier. Only lines logged on the request thread carry one.

Returns:

  • LogTailResponse – The matching lines together with the state of the read.

Raises:

  • HTTPStatusError – If the request fails, for example because a filter value is not valid, the buffer is disabled or the user is not authorized.
  • LogsReadError – If the answer is not a log buffer response.

logger¤

logger = logging.getLogger(f'{self._client.logger.name}.{self.__class__.__name__}')

Comments