Logger

Deprecated since version 0.4.0.

Warning

The Logger interface has been deprecated beginning with release 0.4.0. Please use the Publisher interface instead, using the only_raw: False or send_raw: False Publisher configuration options. For example, if you used the following configuration for the Logger:

logger:
  kafka:
    send_raw: true

The configuration must be updated to:

Using only_raw is recommended to ensure that the Publisher will be used only for this exact purpose. However, the user can decide what is the most suitable for their use case.

The logger subsystem uses the modules from the publisher pluggable subsystem to send partially parsed syslog messages. The configuration options are the same as for the publisher referenced – see the Available publishers and their options. It can be used together with the publisher system in such a way the publisher externalizes the fully processed objects and the clients can subscribe and collect them, while the logger submits the partially parsed messages. This is ideal for logging these unprocessed messages, hence the logger name.

This subsystem is by default disabled and it cannot be configured from the command line, but only from the configuration file. Besides the publisher name to be specified, it also requires to configure at least one set one of the options below:

send_raw

If this option is set, all processed syslog messages, even ones that have not matched a configured error, will be output via the specified transport. This can be used to forward to log server for storage.

Example:

logger:
  kafka:
    send_raw: true

send_unknown

If this option is set, all processed syslog messages, even ones that have not matched a certain operating system, will be output via the specified transport. This can be used to forward to log server for storage.

Example:

logger:
  zmq:
    send_unknown: true