Add documentation about the usage of api-no-pipline

This feature has implemented a different mechanism when posting samples
via API, the related configuration and API usage (a bit different than
before) should be described in doc.

Partially implements: blueprint api-no-pipeline

Change-Id: I8e29aab749aba6754b8dca4356bc0e44c673b9a7
This commit is contained in:
liu-sheng 2015-06-25 11:58:36 +08:00 committed by LiuSheng
parent 1d876f9fbb
commit fa713674a2

View File

@ -667,6 +667,17 @@ Fields *source*, *timestamp*, *project_id* and *user_id* are automatically
added if not present in the samples. Field *message_id* is not taken into
account if present and an internal value will be set.
By default, samples posted via API will be placed on the notification bus and
processed by the notification agent.
To avoid re-queuing the data, samples posted via API can be stored directly to
the storage backend verbatim by specifying a boolean flag 'direct' in the
request URL, like this::
POST /v2/meters/ram_util?direct=True
Samples posted this way will bypass pipeline processing.
Here is an example showing how to add a sample for a *ram_util* meter (already
existing or not)::