Common Message Format

Heka turns the incoming data into Heka messages [1] with a well-defined format which is described below.

  • Timestamp (number), the timestamp of the message (in nanoseconds since the Epoch).
  • Logger (string), the datasource from the Heka’s standpoint.
  • Type (string), the type of message.
  • Hostname (string), the name of the host that emitted the message.
  • Severity (number), severity level as defined by the Syslog RFC 5424.
  • Payload (string), the input data in most cases.
  • Pid (number), the Process ID that generated the message.
  • Fields, array of Field structures (see below).

Field Format

Every message (either originating from logs, metrics or notifications) is populated with a set of predefined fields:

Attributes in bold are always present in the messages while attributes in italic are optional.

  • deployment_mode (string), the deployment mode of the Fuel environment (either ‘multinode’ or ‘ha_compact’).
  • deployment_id (number), the deployment identifier of the Fuel environment.
  • openstack_region (string), the name of the OpenStack region.
  • openstack_release (string), the name of the OpenStack release.
  • openstack_roles (string), a comma-separated list of the node’s roles (eg ‘controller’, ‘compute,cinder’).
  • environment_label (string), the label assigned to the OpenStack environment.

Note

All date/time fields represented as string are formatted according to the RFC3339 document.

[1]Heka message structure