===========================
Octavia HAProxy Amphora API
===========================

Introduction
============
This document describes the API interface between the reference haproxy driver
and its corresponding haproxy-based amphorae.

Octavia reference haproxy amphorae use a web service API for configuration and
control. This API should be secured through the use of TLS encryption as well
as bi-directional verification of client- and server-side certificates. (The
exact process for generating and distributing these certificates should be
covered in another document.)

In addition to the web service configuration and control interface, the
amphorae may use an HMAC-signed UDP protocol for communicating regular, less-
vital information to the controller (ex. statistics updates and health checks).
Information on this will also be covered in another document.

If a given loadbalancer is being serviced by multiple haproxy amphorae at the
same time, configuration and control actions should be made on all these
amphorae at approximately the same time. (Amphorae do not communicate directly
with each other, except in an active-standby topology, and then this
communication is limited to fail-over protocols.)

.. contents::

Versioning
----------
All Octavia APIs (including internal APIs like this one) are versioned. For the
purposes of this document, the initial version of this API shall be 1.0.

Response codes
--------------
Typical response codes are:

* 200 OK - Operation was completed as requested.
* 201 Created - Operation successfully resulted in the creation / processing
  of a file.
* 202 Accepted - Command was accepted but is not completed. (Note that this is
  used for asynchronous processing.)
* 400 Bad Request - API handler was unable to complete request.
* 401 Unauthorized - Authentication of the client certificate failed.
* 404 Not Found - The requested file was not found.
* 500 Internal Server Error - Usually indicates a permissions problem
* 503 Service Unavailable - Usually indicates a change to a listener was
  attempted during a transition of amphora topology.

A note about storing state
--------------------------
In the below API, it will become apparent that at times the amphora will need
to be aware of the state of things (topology-wise, or simply in terms running
processes on the amphora). When it comes to storing or gathering this data, we
should generally prefer to try to resolve these concerns in the following
order. Note also that not every kind of state data will use all of the steps in
this list:

1. Get state information by querying running processes (ex. parsing haproxy
   status page or querying iptables counters, etc.)
2. Get state by consulting on-disk cache generated by querying running
   processes. (In the case where state information is relatively expensive to
   collect-- eg. package version listings.)
3. Get state by consulting stored configuration data as sent by the controller.
   (ex. amphora topology, haproxy configuration or TLS certificate data)
4. Get state by querying a controller API (not described here).

In no case should the amphora assume it ever has direct access to the Octavia
database. Also, sensitive data (like TLS certificates) should be stored in
a secure way (ex. memory filesystem).

API
===

Get amphora info
----------------
* **URL:** /info
* **Method:** GET
* **URL params:** none
* **Data params:** none
* **Success Response:**

  * Code: 200

    * Content: JSON formatted listing of several basic amphora data.

* **Error Response:**

  * none

JSON Response attributes:

* *hostname* - amphora hostname
* *uuid* - amphora UUID
* *haproxy_version* - Version of the haproxy installed
* *api_version* - Version of haproxy amphora API in use

**Notes:** The data in this request is used by the controller for determining
the amphora and API version numbers.

It's also worth noting that this is the only API command that doesn't have a
version string prepended to it.

**Examples:**

* Success code 200:

::

  {
    'hostname': 'octavia-haproxy-img-00328.local',
    'uuid': '6e2bc8a0-2548-4fb7-a5f0-fb1ef4a696ce',
    'haproxy_version': '1.5.11',
    'api_version': '0.1',
  }

Get amphora details
-------------------

* **URL:** /1.0/details
* **Method:** GET
* **URL params:** none
* **Data params:** none
* **Success Response:**

  * Code: 200

    * Content: JSON formatted listing of various amphora statistics.

* **Error Response:**

  * none

JSON Response attributes:

* *hostname* - amphora hostname
* *uuid* - amphora UUID
* *haproxy_version* - Version of the haproxy installed
* *api_version* - Version of haproxy amphora API/agent in use
* *network_tx* - Current total outbound bandwidth in bytes/sec (30-second
  snapshot)
* *network_rx* - Current total inbound bandwidth in bytes/sec (30-second
  snapshot)
* *active* - Boolean (is amphora in an "active" role?)
* *haproxy_count* - Number of running haproxy processes
* *cpu* - list of percent CPU usage broken down into:

  * total
  * user
  * system
  * soft_irq

* *memory* - memory usage in kilobytes broken down into:

  * total
  * free
  * available
  * buffers
  * cached
  * swap_used
  * shared
  * slab
  * committed_as

* *disk* - disk usage in kilobytes for root filesystem, listed as:

  * used
  * available

* *load* - System load (list)
* *topology* - One of SINGLE, ACTIVE-STANDBY, ACTIVE-ACTIVE
* *topology_status* - One of OK, TOPOLOGY-CHANGE
* *listeners* - list of listener UUIDs being serviced by this amphora
* *packages* - list of load-balancing related packages installed with versions
  (eg. OpenSSL, haproxy, nginx, etc.)

**Notes:** The data in this request is meant to provide intelligence for an
auto-scaling orchestration controller (heat) in order to determine whether
additional (or fewer) virtual amphorae are necessary to handle load. As such,
we may add additional parameters to the JSON listing above if they prove to be
useful for making these decisions.

The data in this request is also used by the controller for determining overall
health of the amphora, currently-configured topology and role, etc.

**Examples**

* Success code 200:

::

  {
    'hostname': 'octavia-haproxy-img-00328.local',
    'uuid': '6e2bc8a0-2548-4fb7-a5f0-fb1ef4a696ce',
    'haproxy_version': '1.5.11',
    'api_version': '0.1',
    'networks': {
        'eth0': {
            'network_tx': 3300138,
            'network_rx': 982001, }}
    'active': 'TRUE',
    'haproxy_count': 3,
    'cpu':{
      'total': 0.43,
      'user': 0.30,
      'system': 0.05,
      'soft_irq': 0.08,
    },
    'memory':{
      'total': 4087402,
      'free': 760656,
      'available': 2655901,
      'buffers': 90980,
      'cached': 1830143,
      'swap_used': 943,
      'shared': 105792,
      'slab': 158819,
      'committed_as': 2643480,
    },
    'disk':{
      'used': 1234567,
      'available': 5242880,
    },
    'load': [0.50, 0.45, 0.47],
    'tolopogy': 'SINGLE',
    'topology_status': 'OK',
    'listeners':[
      '02d0da8d-fc65-4bc4-bc46-95cadb2315d2',
      '98e706a7-d22c-422f-9632-499fd83e12c0',
    ],
    'packages':[
      {'haproxy': '1.5.1'},
      {'bash': '4.3.23'},
      {'lighttpd': '1.4.33-1'},
      {'openssl': '1.0.1f'},
      <cut for brevity>
    ],
   }

Get interface
-------------

* **URL:** /1.0/interface/*:ip*
* **Method:** GET
* **URL params:**

     * *:ip* = the ip address to find the interface name

* **Data params:** none
* **Success Response:**

  * Code: 200

    * Content: OK
    * Content: JSON formatted interface

* **Error Response:**

  * Code: 400

    * Content: Bad IP address version

  * Code: 404

    * Content: Error interface not found for IP address

* **Response:**

| OK
| eth1

**Examples:**

* Success code 200:

::

  GET URL:
  https://octavia-haproxy-img-00328.local/1.0/interface/10.0.0.1

  JSON Response:
      {
        'message': 'OK',
        'interface': 'eth1'
      }


* Error code 404:

::

  GET URL:
  https://octavia-haproxy-img-00328.local/1.0/interface/10.5.0.1

  JSON Response:
      {
        'message': 'Error interface not found for IP address',
      }


* Error code 404:

::

  GET URL:
  https://octavia-haproxy-img-00328.local/1.0/interface/10.6.0.1.1

  JSON Response:
      {
        'message': 'Bad IP address version',
      }


Get all listeners' statuses
---------------------------

* **URL:** /1.0/listeners
* **Method:** GET
* **URL params:** none
* **Data params:** none
* **Success Response:**

  * Code: 200

    * Content: JSON-formatted listing of each listener's status

* **Error Response:**

  * none

JSON Response attributes:

Note that the command will return an array of *all* listeners' statuses. Each
listener status contains the following attributes:

* *status* - One of the operational status: ACTIVE, STOPPED, ERROR -
  future versions might support provisioning status:
  PENDING_CREATE, PENDING_UPDATE, PENDING_DELETE, DELETED
* *uuid* - Listener UUID
* *type* - One of: TCP, HTTP, TERMINATED_HTTPS

**Notes:** Note that this returns a status if: the pid file exists, the stats
socket exists, or an haproxy configuration is present (not just if there is
a valid haproxy configuration).

**Examples**

* Success code 200:

::

  [{
    'status': 'ACTIVE',
    'uuid': 'e2dfddc0-5b9e-11e4-8ed6-0800200c9a66',
    'type': 'HTTP',
   },
   {
    'status': 'STOPPED',
    'uuid': '19d45130-5b9f-11e4-8ed6-0800200c9a66',
    'type': 'TERMINATED_HTTPS',
   }]

Start or Stop a load balancer
-----------------------------

* **URL:** /1.0/loadbalancer/*:object_id*/*:action*
* **Method:** PUT
* **URL params:**

  * *:object_id* = Object UUID
  * *:action* = One of: start, stop, reload

* **Data params:** none
* **Success Response:**

  * Code: 202

    * Content: OK
    * *(Also contains preliminary results of attempt to start / stop / soft \
      restart (reload) the haproxy daemon)*

* **Error Response:**

  * Code: 400

    * Content: Invalid request

  * Code: 404

    * Content: Listener Not Found

  * Code: 500

    * Content: Error starting / stopping / reload_config haproxy
    * *(Also contains error output from attempt to start / stop / soft \
      restart (reload) haproxy)*

  * Code: 503

    * Content: Topology transition in progress

* **Response:**

| OK
| Configuration file is valid
| haproxy daemon for 85e2111b-29c4-44be-94f3-e72045805801 started (pid 32428)

**Examples:**

* Success code 201:

::

  PUT URL:
  https://octavia-haproxy-img-00328.local/1.0/loadbalancer/85e2111b-29c4-44be-94f3-e72045805801/start

  JSON Response:
  {
    'message': 'OK',
    'details': 'Configuration file is valid\nhaproxy daemon for 85e2111b-29c4-44be-94f3-e72045805801 started',
  }

* Error code 400:

::

  PUT URL:
  https://octavia-haproxy-img-00328.local/1.0/loadbalancer/85e2111b-29c4-44be-94f3-e72045805801/BAD_TEST_DATA

  JSON Response:
  {
    'message': 'Invalid Request',
    'details': 'Unknown action: BAD_TEST_DATA',
  }

* Error code 404:

::

  PUT URL:
  https://octavia-haproxy-img-00328.local/1.0/loadbalancer/04bff5c3-5862-4a13-b9e3-9b440d0ed50a/stop

  JSON Response:
  {
    'message': 'Listener Not Found',
    'details': 'No loadbalancer with UUID: 04bff5c3-5862-4a13-b9e3-9b440d0ed50a',
  }

* Error code 500:

::

  PUT URL:
  https://octavia-haproxy-img-00328.local/1.0/loadbalancer/85e2111b-29c4-44be-94f3-e72045805801/stop

  Response:
  {
    'message': 'Error stopping haproxy',
    'details': 'haproxy process with PID 3352 not found',
  }

* Error code 503:

::

  Response:
  {
    'message': 'Topology transition in progress',
  }

Delete a listener
-----------------

* **URL:** /1.0/listeners/*:listener*
* **Method:** DELETE
* **URL params:**

  * *:listener* = Listener UUID

* **Data params:** none
* **Success Response:**

  * Code: 200

    * Content: OK

* **Error Response:**

  * Code: 404

    * Content: Not Found

  * Code: 503

    * Content: Topology transition in progress

* **Response:**

| OK

* **Implied actions:**

  * Stop listener
  * Delete IPs, iptables accounting rules, etc. from this amphora if they're no
    longer in use.
  * Clean up listener configuration directory.
  * Delete listener's SSL certificates
  * Clean up logs (ship final logs to logging destination if configured)
  * Clean up stats socket.

**Examples**

* Success code 200:

::

  DELETE URL:
  https://octavia-haproxy-img-00328.local/1.0/listeners/04bff5c3-5862-4a13-b9e3-9b440d0ed50a

  JSON Response:
  {
    'message': 'OK'
  }

* Error code 404:

::

  DELETE URL:
  https://octavia-haproxy-img-00328.local/1.0/listeners/04bff5c3-5862-4a13-b9e3-9b440d0ed50a

  JSON Response:
  {
    'message': 'Listener Not Found',
    'details': 'No listener with UUID: 04bff5c3-5862-4a13-b9e3-9b440d0ed50a',
  }

* Error code 503:

::

  Response:
  {
    'message': 'Topology transition in progress',
  }

Upload SSL certificate PEM file
-------------------------------

* **URL:** /1.0/loadbalancer/*:loadbalancer_id*/certificates/*:filename.pem*
* **Method:** PUT
* **URL params:**

  * *:loadbalancer_id* = Load balancer UUID
  * *:filename* = PEM filename (see notes below for naming convention)

* **Data params:** Certificate data. (PEM file should be a concatenation of
  unencrypted RSA key, certificate and chain, in that order)
* **Success Response:**

  * Code: 201

    * Content: OK

* **Error Response:**

  * Code: 400

    * Content: No certificate found

  * Code: 400

    * Content: No RSA key found

  * Code: 400

    * Content: Certificate and key do not match

  * Code: 404

    * Content: Not Found

  * Code: 503

    * Content: Topology transition in progress

* **Response:**

| OK

**Notes:**
* filename.pem should match the primary CN for which the
certificate is valid. All-caps WILDCARD should be used to replace an asterisk
in a wildcard certificate (eg. a CN of '\*.example.com' should have a filename
of 'WILDCARD.example.com.pem'). Filenames must also have the .pem extension.
* In order for the new certificate to become effective the haproxy needs to be
explicitly restarted

**Examples:**

* Success code 201:

::

  PUT URI:
  https://octavia-haproxy-img-00328.local/1.0/loadbalancer/85e2111b-29c4-44be-94f3-e72045805801/certificates/www.example.com.pem
  (Put data should contain the certificate information, concatenated as
  described above)

  JSON Response:
  {
    'message': 'OK'
  }

* Error code 400:

::

  PUT URI:
  https://octavia-haproxy-img-00328.local/1.0/loadbalancer/85e2111b-29c4-44be-94f3-e72045805801/certificates/www.example.com.pem
  (If PUT data does not contain a certificate)

  JSON Response:
  {
    'message': 'No certificate found'
  }

* Error code 400:

::

  PUT URI:
  https://octavia-haproxy-img-00328.local/1.0/loadbalancer/85e2111b-29c4-44be-94f3-e72045805801/certificates/www.example.com.pem
  (If PUT data does not contain an RSA key)

  JSON Response:
  {
    'message': 'No RSA key found'
  }

* Error code 400:

::

  PUT URI:
  https://octavia-haproxy-img-00328.local/1.0/loadbalancer/85e2111b-29c4-44be-94f3-e72045805801/certificates/www.example.com.pem
  (If the first certificate and the RSA key do not have the same modulus.)

  JSON Response:
  {
    'message': 'Certificate and key do not match'
  }

* Error code 404:

::

  PUT URI:
  https://octavia-haproxy-img-00328.local/1.0/loadbalancer/85e2111b-29c4-44be-94f3-e72045805801/certificates/www.example.com.pem

  JSON Response:
  {
    'message': 'Listener Not Found',
    'details': 'No loadbalancer with UUID: 04bff5c3-5862-4a13-b9e3-9b440d0ed50a',
  }

* Error code 503:

::

  Response:
  {
    'message': 'Topology transition in progress',
  }

Get SSL certificate md5sum
--------------------------

* **URL:** /1.0/loadbalancer/*:loadbalancer_id*/certificates/*:filename.pem*
* **Method:** GET
* **URL params:**

  * *:loadbalancer_id* = Load balancer UUID
  * *:filename* = PEM filename (see notes below for naming convention)

* **Data params:** none
* **Success Response:**

  * Code: 200

    * Content: PEM file md5sum

* **Error Response:**

  * Code: 404

    * Content: Not Found

* **Response:**

| <certificate PEM file md5 sum>

* **Implied actions:** none

**Notes:** The md5sum is the sum from the raw certificate data as stored on
the amphora (which will usually include the RSA key, certificate and chain
concatenated together). Note that we don't return any actual raw certificate
data as the controller should already know this information, and unnecessarily
disclosing it over the wire from the amphora is a security risk.

**Examples:**

* Success code 200:

::

  JSON response:
  {
    'md5sum': 'd8f6629d5e3c6852fa764fb3f04f2ffd',
  }

* Error code 404:

::

    JSON Response:
      {
        'message': 'Listener Not Found',
        'details': 'No loadbalancer with UUID: 04bff5c3-5862-4a13-b9e3-9b440d0ed50a',
      }

* Error code 404:

::

    JSON Response:
      {
        'message': 'Certificate Not Found',
        'details': 'No certificate with file name: www.example.com.pem',
      }

Delete SSL certificate PEM file
-------------------------------

* **URL:** /1.0/loadbalancer/*:loadbalancer_id*/certificates/*:filename.pem*
* **Method:** DELETE
* **URL params:**

  * *:loadbalancer_id* = Load balancer UUID
  * *:filename* = PEM filename (see notes below for naming convention)

* **Data params:** none
* **Success Response:**

  * Code: 200

    * Content: OK

* **Error Response:**

  * Code: 404

    * Content: Not found

  * Code: 503

    * Content: Topology transition in progress

* **Implied actions:**

  * Clean up listener configuration directory if it's now empty.

**Examples:**

* Success code 200:

::

  DELETE URL:
  https://octavia-haproxy-img-00328.local/1.0/loadbalancer/85e2111b-29c4-44be-94f3-e72045805801/certificates/www.example.com.pem

  JSON Response:
  {
    'message': 'OK'
  }

* Error code 404:

::

  DELETE URL:
  https://octavia-haproxy-img-00328.local/1.0/loadbalancer/85e2111b-29c4-44be-94f3-e72045805801/certificates/www.example.com.pem

 JSON Response:
      {
        'message': 'Certificate Not Found',
        'details': 'No certificate with file name: www.example.com.pem',
      }

* Error code 503:

::

  Response:
  {
    'message': 'Topology transition in progress',
  }

Upload load balancer haproxy configuration
------------------------------------------

* **URL:** /1.0/loadbalancer/*:amphora_id*/*:loadbalancer_id*/haproxy
* **Method:** PUT
* **URL params:**

  * *:loadbalancer_id* = Load Balancer UUID
  * *:amphora_id* = Amphora UUID

* **Data params:** haproxy configuration file for the listener
* **Success Response:**

  * Code: 201

    * Content: OK

* **Error Response:**

  * Code: 400

    * Content: Invalid configuration
    * *(Also includes error output from configuration check command)*

  * Code: 503

    * Content: Topology transition in progress

* **Response:**

| OK
| Configuration file is valid

* **Implied actions:**

  * Do a syntax check on haproxy configuration file prior to an attempt to
    run it.
  * Add resources needed for stats, logs, and connectivity

**Notes:** The uploaded configuration file should be a complete and
syntactically-correct haproxy config. The amphora does not have intelligence
to generate these itself and has only rudimentary ability to parse certain
features out of the configuration file (like bind addresses and ports for
purposes of setting up stats, and specially
formatted comments meant to indicate pools and members that will be parsed
out of the haproxy daemon status interface for tracking health and stats).

**Examples:**

* Success code 201:

::

  PUT URL:
  https://octavia-haproxy-img-00328.local/1.0/loadbalancer/d459b1c8-54b0-4030-9bec-4f449e73b1ef/85e2111b-29c4-44be-94f3-e72045805801/haproxy
  (Upload PUT data should be a raw haproxy.conf file.)

  JSON Response:
  {
    'message': 'OK'
  }

* Error code 400:

::

  JSON Response:
  {
    'message': 'Invalid request',
    'details': '[ALERT] 300/013045 (28236) : parsing [haproxy.cfg:4]: unknown keyword 'BAD_LINE' out of section.\n[ALERT] 300/013045 (28236) : Error(s) found in configuration file : haproxy.cfg\n[ALERT] 300/013045 (28236) : Fatal errors found in configuration.',
  }

* Error code 503:

::

  Response:
  {
    'message': 'Topology transition in progress',
  }

Get loadbalancer haproxy configuration
--------------------------------------

* **URL:** /1.0/loadbalancer/*:loadbalancer_id*/haproxy
* **Method:** GET
* **URL params:**

  * *:loadbalancer_id* = Load balancer UUID

* **Data params:** none
* **Success Response:**

  * Code: 200

    * Content: haproxy configuration file for the listener

* **Error Response:**

  * Code: 404

    * Content: Not found

* **Response:**

| # Config file for 85e2111b-29c4-44be-94f3-e72045805801
| (cut for brevity)

* **Implied actions:** none

**Examples:**

* Success code 200:

::

  GET URL:
  https://octavia-haproxy-img-00328.local/1.0/loadbalancer/85e2111b-29c4-44be-94f3-e72045805801/haproxy

  Response is the raw haproxy.cfg:

  # Config file for 85e2111b-29c4-44be-94f3-e72045805801
  (cut for brevity)

* Error code 404:

::

    JSON Response:
      {
        'message': 'Loadbalancer Not Found',
        'details': 'No loadbalancer with UUID: 04bff5c3-5862-4a13-b9e3-9b440d0ed50a',
      }

Plug VIP
--------

* **URL:** /1.0/plug/vip/*:ip*
* **Method:** Post
* **URL params:**

     * *:ip* = the vip's ip address

* **Data params:**

 * *subnet_cidr*: The vip subnet in cidr notation
 * *gateway*: The vip subnet gateway address
 * *mac_address*: The mac address of the interface to plug

* **Success Response:**

  * Code: 202

    * Content: OK

* **Error Response:**
  * Code: 400

    * Content: Invalid IP
    * Content: Invalid subnet information

  * Code: 404

    * Content: No suitable network interface found

  * Code: 500

    * Content: Error plugging VIP
    * (Also contains error output from the ip up command)

  * Code: 503

    * Content: Topology transition in progress

* **Response:**

| OK
| VIP <vip> ip plugged on interface <interface>

* **Implied actions:**

  * Look for an interface marked as down (recently added port)
  * Assign VIP
  * Bring that interface up

**Examples:**

* Success code 202:

::

  POST URL:
  https://octavia-haproxy-img-00328.local/1.0/plug/vip/203.0.113.2

  JSON POST parameters:
  {
    'subnet_cidr': '203.0.113.0/24',
    'gateway': '203.0.113.1',
    'mac_address': '78:31:c1:ce:0b:3c'
  }

  JSON Response:
      {
        'message': 'OK',
        'details': 'VIP 203.0.113.2 plugged on interface eth1'
      }

* Error code 400:

::

    JSON Response:
      {
        'message': 'Invalid VIP',
      }

* Error code 404:

::

    JSON Response:
      {
        'message': 'No suitable network interface found',
      }


Plug Network
------------

* **URL:** /1.0/plug/network/
* **Method:** POST
* **URL params:** none

* **Data params:**

 * *mac_address*: The mac address of the interface to plug

* **Success Response:**

  * Code: 202

    * Content: OK

* **Error Response:**

  * Code: 404

    * Content: No suitable network interface found

  * Code: 500

    * Content: Error plugging Port
    * (Also contains error output from the ip up command)

  * Code: 503

    * Content: Topology transition in progress

* **Response:**

| OK
| Plugged interface <interface>

**Examples:**

* Success code 202:

::

  POST URL:
  https://octavia-haproxy-img-00328.local/1.0/plug/network/

  JSON POST parameters:
  {
    'mac_address': '78:31:c1:ce:0b:3c'
  }

  JSON Response:
      {
        'message': 'OK',
        'details': 'Plugged interface eth1'
      }


* Error code 404:

::

    JSON Response:
      {
        'message': 'No suitable network interface found',
      }


Upload SSL server certificate PEM file for Controller Communication
-------------------------------------------------------------------

* **URL:** /1.0/certificate
* **Method:** PUT

* **Data params:** Certificate data. (PEM file should be a concatenation of
  unencrypted RSA key, certificate and chain, in that order)
* **Success Response:**

  * Code: 202

    * Content: OK

* **Error Response:**

  * Code: 400

    * Content: No certificate found

  * Code: 400

    * Content: No RSA key found

  * Code: 400

    * Content: Certificate and key do not match


* **Response:**

| OK

**Notes:**
Since certificates might be valid for a time smaller than the amphora is in
existence this add a way to rotate them. Once the certificate is uploaded the
agent is being recycled so depending on the implementation the service might
not be available for some time.

**Examples:**

* Success code 202:

::

  PUT URI:
  https://octavia-haproxy-img-00328.local/1.0/certificate
  (Put data should contain the certificate information, concatenated as
  described above)

  JSON Response:
  {
    'message': 'OK'
  }

* Error code 400:

::

  PUT URI:
  https://octavia-haproxy-img-00328.local/1.0/certificates
  (If PUT data does not contain a certificate)

  JSON Response:
  {
    'message': 'No certificate found'
  }

* Error code 400:

::

  PUT URI:
  https://octavia-haproxy-img-00328.local/1.0/certificate
  (If PUT data does not contain an RSA key)

  JSON Response:
  {
    'message': 'No RSA key found'
  }

* Error code 400:

::

  PUT URI:
  https://octavia-haproxy-img-00328.local/1.0/certificate
  (If the first certificate and the RSA key do not have the same modulus.)

  JSON Response:
  {
    'message': 'Certificate and key do not match'
  }


Upload keepalived configuration
-------------------------------

* **URL:** /1.0/vrrp/upload
* **Method:** PUT
* **URL params:** none
* **Data params:** none
* **Success Response:**

  * Code: 200

    * Content: OK

* **Error Response:**

  * Code: 500

    * Content: Failed to upload keepalived configuration.

* **Response:**

OK

**Examples:**

* Success code 200:

::

  PUT URI:
  https://octavia-haproxy-img-00328.local/1.0/vrrp/upload

  JSON Response:
  {
    'message': 'OK'
  }


Start, Stop, or Reload keepalived
---------------------------------

* **URL:** /1.0/vrrp/*:action*
* **Method:** PUT
* **URL params:**

  * *:action* = One of: start, stop, reload

* **Data params:** none
* **Success Response:**

  * Code: 202

    * Content: OK

* **Error Response:**

  * Code: 400

    * Content: Invalid Request

  * Code: 500

    * Content: Failed to start / stop / reload keepalived service:
    * *(Also contains error output from attempt to start / stop / \
      reload keepalived)*

* **Response:**

| OK
| keepalived started

**Examples:**

* Success code 202:

::

  PUT URL:
  https://octavia-haproxy-img-00328.local/1.0/vrrp/start

  JSON Response:
  {
    'message': 'OK',
    'details': 'keepalived started',
  }

* Error code: 400

::

  PUT URL:
  https://octavia-haproxy-img-00328.local/1.0/vrrp/BAD_TEST_DATA

  JSON Response:
  {
    'message': 'Invalid Request',
    'details': 'Unknown action: BAD_TEST_DATA',
  }

* Error code: 500

::

  PUT URL:
  https://octavia-haproxy-img-00328.local/1.0/vrrp/stop

  JSON Response:
  {
    'message': 'Failed to stop keepalived service: keeepalived process with PID 3352 not found',
    'details': 'keeepalived process with PID 3352 not found',
  }

Update the amphora agent configuration
--------------------------------------

* **URL:** /1.0/config
* **Method:** PUT

* **Data params:** A amphora-agent configuration file
* **Success Response:**

  * Code: 202

    * Content: OK

* **Error Response:**

  * Code: 500

    * message: Unable to update amphora-agent configuration.
    * details: *(The exception details)*

* **Response:**

| OK

* **Implied actions:**

  * The running amphora-agent configuration file is mutated.

**Notes:** Only options that are marked mutable in the oslo configuration
will be updated.

**Examples:**

* Success code 202:

::

  PUT URL:
  https://octavia-haproxy-img-00328.local/1.0/config
  (Upload PUT data should be a raw amphora-agent.conf file.)

  JSON Response:
  {
    'message': 'OK'
  }

* Error code 500:

::

  JSON Response:
  {
    'message': 'Unable to update amphora-agent configuration.',
    'details': *(The exception output)*,
  }