treeherder/docs/submitting_data.rst

524 строки
18 KiB
ReStructuredText

Submitting Data
===============
To submit your test data to Treeherder, you have two options:
1. :ref:`submitting-using-pulse`
This is the new process Task Cluster is using to submit data to Treeherder.
There is a `Pulse Job Schema`_ to validate your payload against to ensure it will
be accepted. In this case, you create your own `Pulse`_ exchange and publish
to it. To get Treeherder to receive your data, you would create a bug to
have your Exchange added to Treeherder's config. All Treeherder instances
can subscribe to get your data, as can local dev instances for testing.
While it is beyond the scope of this document to explain how `Pulse`_ and
RabbitMQ work, we encourage you to read more about this technology on
its Wiki page.
2. :ref:`submitting-using-python-client`
This is historically how projects and users have submitted data to Treeherder.
This requires getting Hawk credentials approved by a Treeherder Admin.
There is a client library to help make this easier. However, there is no
schema to validate the payload against. But using the client to build your
payload will help you get it in the accepted form. Your data only goes to
the host you send it to. Dev instances can not subscribe to this data.
If you are establishing a new repository with Treeherder, then you will need to
do one of the following:
1. For GitHub repos: :ref:`add-github-repo`
2. For Mercurial repos: :ref:`add-hg-repo`
.. _submitting-using-pulse:
Using Pulse
-----------
To submit via a Pulse exchange, these are the steps you will need to follow:
1. Format your data
^^^^^^^^^^^^^^^^^^^
You should format your job data according to the `Pulse Job Schema`_,
which describes the various properties of a job: whether it passed or failed,
job group/type symbol, description, log information, etc.
You are responsible for validating your data prior to publishing it onto your
exchange, or Treeherder may reject it.
2. Create your Exchange
^^^^^^^^^^^^^^^^^^^^^^^
With `Pulse Guardian`_, you need to create your Pulse User in order to
create your own Queues and Exchanges. There is no mechanism to create an
Exchange in the Pulse Guardian UI itself, however. You will need to create
your exchange in your submitting code. There are a few options available
for that:
1. `MozillaPulse`_
2. `Kombu`_
3. Or any RabbitMQ package of your choice
To test publishing your data to your Exchange, you can use the Treeherder
management command `publish_to_pulse`_. This is also a very simple example
of a Pulse publisher using Kombu that you can use to learn to write your own
publisher.
3. Register with Treeherder
^^^^^^^^^^^^^^^^^^^^^^^^^^^
Once you have successfully tested a round-trip through your Pulse exchange to
your development instance, you are ready to have Treeherder receive your data.
Treeherder has to know about your exchange and which routing keys to use in
order to load your jobs.
Submit a `Treeherder bug`_ with the following information::
{
"exchange": "exchange/my-pulse-user/v1/jobs",
"destinations": [
'treeherder'
],
"projects": [
'mozilla-inbound._'
],
},
Treeherder will bind to the exchange looking for all combinations of routing
keys from ``destinations`` and ``projects`` listed above. For example with
the above config, we will only load jobs with routing keys of
``treeherder.mozilla-inbound._``
If you want all jobs from your exchange to be loaded, you could simplify the
config by having values::
"destinations": [
'#'
],
"projects": [
'#'
],
If you want one config to go to Treeherder Staging and a different one to go
to Production, please specify that in the bug. You could use the same exchange
with different routing key settings, or two separate exchanges. The choice is
yours.
4. Publish jobs to your Exchange
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Once the above config is set on Treeherder, you can begin publishing jobs
to your Exchange and they will start showing in Treeherder.
You will no longer need any special credentials. You publish messages to the
Exchange YOU own. Treeherder is now just listening to it.
.. _submitting-using-python-client:
Using the Python Client
-----------------------
There are two types of data structures you can submit with the :ref:`Python client
<python-client>`: job and push collections. The client provides methods
for building a data structure that treeherder will accept. Data
structures can be extended with new properties as needed, there is a
minimal validation protocol applied that confirms the bare minimum
parts of the structures are defined.
See the :ref:`Python client <python-client>` section for how to control
which Treeherder instance will be accessed by the client.
Authentication is covered :ref:`here <authentication>`.
Job Collections
^^^^^^^^^^^^^^^
Job collections can contain test results from any kind of test. The
`revision` provided should match the associated `revision` in the
push structure. The `revision` is the top-most revision in the push.
The `job_guid` provided can be any unique string of 50
characters at most. A job collection has the following data structure.
.. code-block:: python
[
{
'project': 'mozilla-inbound',
'revision': '4317d9e5759d58852485a7a808095a44bc806e19',
'job': {
'job_guid': 'd22c74d4aa6d2a1dcba96d95dccbd5fdca70cf33',
'product_name': 'spidermonkey',
'reason': 'scheduler',
'who': 'spidermonkey_info__mozilla-inbound-warnaserr',
'desc': 'Linux x86-64 mozilla-inbound spidermonkey_info-warnaserr build',
'name': 'SpiderMonkey --enable-sm-fail-on-warnings Build',
# The symbol representing the job displayed in
# treeherder.allizom.org
'job_symbol': 'e',
# The symbol representing the job group in
# treeherder.allizom.org
'group_symbol': 'SM',
'group_name': 'SpiderMonkey',
'submit_timestamp': 1387221298,
'start_timestamp': 1387221345,
'end_timestamp': 1387222817,
'state': 'completed',
'result': 'success',
'machine': 'bld-linux64-ec2-104',
'build_platform': {
'platform':'linux64', 'os_name': 'linux', 'architecture': 'x86_64'
},
'machine_platform': {
'platform': 'linux64', 'os_name': 'linux', 'architecture': 'x86_64'
},
'option_collection': {'opt': True},
# jobs can belong to different tiers
# setting the tier here will determine which tier the job
# belongs to. However, if a job is set as Tier of 1, but
# belongs to the Tier 2 profile on the server, it will still
# be saved as Tier 2.
'tier': 2,
# the ``name`` of the log can be the default of "buildbot_text"
# however, you can use a custom name. See below.
'log_references': [
{
'url': 'http://ftp.mozilla.org/pub/mozilla.org/spidermonkey/...',
'name': 'buildbot_text'
}
],
# The artifact can contain any kind of structured data associated with a test.
'artifacts': [{
'type': 'json',
'name': '',
'blob': { my json content here}
}],
# List of job guids that were superseded by this job
'superseded': []
},
...
]
see :ref:`custom-log-name` for more info.
Usage
^^^^^
If you want to use `TreeherderJobCollection` to build up the job data
structures to send, do something like this:
.. code-block:: python
from thclient import (TreeherderClient, TreeherderClientError,
TreeherderJobCollection)
tjc = TreeherderJobCollection()
for data in dataset:
tj = tjc.get_job()
tj.add_revision( data['revision'] )
tj.add_project( data['project'] )
tj.add_coalesced_guid( data['superseded'] )
tj.add_job_guid( data['job_guid'] )
tj.add_job_name( data['name'] )
tj.add_job_symbol( data['job_symbol'] )
tj.add_group_name( data['group_name'] )
tj.add_group_symbol( data['group_symbol'] )
tj.add_description( data['desc'] )
tj.add_product_name( data['product_name'] )
tj.add_state( data['state'] )
tj.add_result( data['result'] )
tj.add_reason( data['reason'] )
tj.add_who( data['who'] )
tj.add_tier( 1 )
tj.add_submit_timestamp( data['submit_timestamp'] )
tj.add_start_timestamp( data['start_timestamp'] )
tj.add_end_timestamp( data['end_timestamp'] )
tj.add_machine( data['machine'] )
tj.add_build_info(
data['build']['os_name'], data['build']['platform'], data['build']['architecture']
)
tj.add_machine_info(
data['machine']['os_name'], data['machine']['platform'], data['machine']['architecture']
)
tj.add_option_collection( data['option_collection'] )
tj.add_log_reference( 'buildbot_text', data['log_reference'] )
# data['artifact'] is a list of artifacts
for artifact_data in data['artifact']:
tj.add_artifact(
artifact_data['name'], artifact_data['type'], artifact_data['blob']
)
tjc.add(tj)
client = TreeherderClient(client_id='hawk_id', secret='hawk_secret')
client.post_collection('mozilla-central', tjc)
If you don't want to use `TreeherderJobCollection` to build up the data structure
to send, build the data structures directly and add them to the collection.
.. code-block:: python
from thclient import TreeherderClient, TreeherderJobCollection
tjc = TreeherderJobCollection()
for job in job_data:
tj = tjc.get_job(job)
# Add any additional data to tj.data here
# add job to collection
tjc.add(tj)
client = TreeherderClient(client_id='hawk_id', secret='hawk_secret')
client.post_collection('mozilla-central', tjc)
Job artifacts format
^^^^^^^^^^^^^^^^^^^^
Artifacts can have name, type and blob. The blob property can contain any
valid data structure accordingly to type attribute. For example if you use
the json type, your blob must be json-serializable to be valid. The name
attribute can be any arbitrary string identifying the artifact. Here is an
example of what a job artifact looks like in the context of a job object:
.. code-block:: python
[
{
'project': 'mozilla-inbound',
'revision': '4317d9e5759d58852485a7a808095a44bc806e19',
'job': {
'job_guid': 'd22c74d4aa6d2a1dcba96d95dccbd5fdca70cf33',
# ...
# other job properties here
# ...
'artifacts': [
{
"type": "json",
"name": "my first artifact",
'blob': {
k1: v1,
k2: v2,
...
}
},
{
'type': 'json',
'name': 'my second artifact',
'blob': {
k1: v1,
k2: v2,
...
}
}
]
}
},
...
]
A special case of job artifact is a "Job Info" artifact. This kind of artifact
will be retrieved by the UI and rendered in the job detail panel. This
is what a Job Info artifact looks like:
.. code-block:: python
{
"blob": {
"job_details": [
{
"url": "https://www.mozilla.org",
"value": "website",
"content_type": "link",
"title": "Mozilla home page"
},
{
"value": "bar",
"content_type": "text",
"title": "Foo"
},
{
"value": "This is <strong>cool</strong>",
"content_type": "raw_html",
"title": "Cool title"
}
],
},
"type": "json",
"name": "Job Info"
}
All the elements in the job_details attribute of this artifact have a
mandatory title attribute and a set of optional attributes depending on
`content_type`. The `content_type` drives the way this kind of artifact
will be rendered. Here are the possible values:
* **Text** - This is the simplest content type you can render and is the one
used by default if the content type specified is not recognised or is missing.
This content type renders as:
.. code-block:: html
<label>{{title}}</label><span>{{value}}</span>
* **Link** - This content type renders as an anchor html tag with the
following format:
.. code-block:: html
{{title}}: <a title="{{value}}" href="{{url}}" target="_blank" rel="noopener">{{value}}</a>
* **Raw Html** - The last resource for when you need to show some formatted
content.
Some Specific Collection POSTing Rules
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Treeherder will detect what data is submitted in the ``TreeherderCollection``
and generate the necessary artifacts accordingly. The outline below describes
what artifacts *Treeherder* will generate depending on what has been submitted.
See :ref:`schema_validation` for more info on validating some specialized JSON
data.
JobCollections
~~~~~~~~~~~~~~
Via the ``/jobs`` endpoint:
1. Submit a Log URL with no ``parse_status`` or ``parse_status`` set to "pending"
* This will generate ``text_log_summary`` and ``Bug suggestions`` artifacts
* Current *Buildbot* workflow
2. Submit a Log URL with ``parse_status`` set to "parsed" and a ``text_log_summary`` artifact
* Will generate a ``Bug suggestions`` artifact only
* Desired future state of *Task Cluster*
3. Submit a Log URL with ``parse_status`` of "parsed", with ``text_log_summary`` and ``Bug suggestions`` artifacts
* Will generate nothing
4. Submit a ``text_log_summary`` artifact
* Will generate a ``Bug suggestions`` artifact if it does not already exist for that job.
5. Submit ``text_log_summary`` and ``Bug suggestions`` artifacts
* Will generate nothing
* This is *Treeherder's* current internal log parser workflow
.. _custom-log-name:
Specifying Custom Log Names
^^^^^^^^^^^^^^^^^^^^^^^^^^^
By default, the Log Viewer expects logs to have the name of ``buildbot_text``
at this time. However, if you are supplying the ``text_log_summary`` artifact
yourself (rather than having it generated for you) you can specify a custom
log name. You must specify the name in two places for this to work.
1. When you add the log reference to the job:
.. code-block:: python
tj.add_log_reference( 'my_custom_log', data['log_reference'] )
2. In the ``text_log_summary`` artifact blob, specify the ``logname`` param.
This artifact is what the Log Viewer uses to find the associated log lines
for viewing.
.. code-block:: python
{
"blob":{
"step_data": {
"steps": [
{
"errors": [ ],
"name": "step",
"started_linenumber": 1,
"finished_linenumber": 1,
"finished": "2015-07-08 06:13:46",
"result": "success",
}
],
"errors_truncated": false
},
"logurl": "https://example.com/mylog.log",
"logname": "my_custom_log"
},
"type": "json",
"id": 10577808,
"name": "text_log_summary",
"job_id": 1774360
}
.. _add-github-repo:
Adding a GitHub Repository
--------------------------
The pushes from GitHub repos come to Treeherder via Pulse. The webhook to enable
this exists in the GitHub group ``mozilla``. (For example, ``github.com/mozilla/treeherder``)
The following steps are required:
1. Create a PR with the new repository information added to the fixtures file:
``treeherder/model/fixtures/repository.json`` (See other entries in that file
for examples of the data to fill.)
2. Open a bug request to enable the webhook that will trigger pulse messages for
every push from your repo. Use the following information:
a. Component: GitHub: Administration
b. Ask to install the https://github.com/apps/taskcluster integration on your repositories
c. List the repositories you want to have access to the integration
d. Answer: Are any of those repositories private?
e. State that this is only to get Pulse messages for integration into Treeherder
.. _Pulse Guardian: https://pulseguardian.mozilla.org/whats_pulse
.. _Pulse: https://wiki.mozilla.org/Auto-tools/Projects/Pulse
.. _Pulse Inspector: https://tools.taskcluster.net/pulse-inspector/
.. _Pulse Job Schema: https://github.com/mozilla/treeherder/blob/master/schemas/pulse-job.yml
.. _Treeherder bug: https://bugzilla.mozilla.org/enter_bug.cgi?component=Treeherder:%20Data%20Ingestion&form_name=enter_bug&op_sys=All&product=Tree%20Management&rep_platform=All
.. _MozillaPulse: https://pypi.python.org/pypi/MozillaPulse
.. _Kombu: https://pypi.python.org/pypi/kombu
.. _publish_to_pulse: https://github.com/mozilla/treeherder/blob/master/treeherder/etl/management/commands/publish_to_pulse.py#L12-L12