Any time you use a REST API, you'll need to be able (for good ones) authenticate and make GET and POST requests. If you're not familiar with GET and POST requests, you may want to freshen up at W3Schools. You'll use your GET and POST methods to send and receive information to and from the array. In this demonstration, we're going to be using the python requests library to make our GET and POST requests to the array.
We're going to use the following information. But first, you need a REST API Token for this demonstration. Here are the steps:
- Navigate to your FlashArray IP: https://10.204.112.109
- Navigate to the Users section: System -> Users -> API Tokens.
- If you don't already have one, create an API token.
Here's the information for reference from our documentation. These IPs and tokens have been changed for security.
|
Hostname |
slc-coz |
|
IP |
|
|
Token |
899fdd4f-50ce-c5bf-f253-29be01c51920 |
|
FlashStache |
The next cell is just our imports and some housekeeping. If you're not familiar with python, "import" is a command that imports a set of commands/scripts you can use that someone else wrote.
In [1]:
import json
import pandas
import requests
import time
import urllib3
# We're using a version that warns us about insecure requests, and we don't want to see that noise.
urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)
%matplotlib inline
Establishing a session with the array
The first thing we need to do is establish a session with the array. This article demonstrates a nifty Python object that manages that session for us. While there are other languages that can manage the sessions for us, our API wrappers are written in Python.
In [2]:
session = requests.Session()
# We're going to use the full URL without any conventional ".format()" type stuff for the demo.
rest_api_url = 'https://slc-m10.slc.purestorage.com/...2-3d634318235d'
# Now we'll submit a POST request to the URL, and don't verify our SSL because it's not set up.
result = session.post(rest_api_url, verify=False)
In [3]:
result.text
Out[3]:
'{"username": "pureuser"}'
Now that you have a session, you can make some API calls. Begin with a simple REST API call to get all the volumes on our array. While there's usually a lot more information, we're using a shortcut in Python to get just the first volume. To get all of them, the command is as follows:
session.get('https://slc-coz.slc.purestorage.com/api/1.12/volume').json()
Adding a [1] at the end to get just the first volume
session.get('https://slc-coz.slc.purestorage.com/api/1.12/volume').json()
In [10]:
session.get('https://slc-coz.slc.purestorage.com/api/1.12/volume').json()[1]
Out [10]:
{'created': '2014-01-23T00:11:00Z',
'name': 'ISO-Repository',
'serial': '4966C24FCE8A51E90001007E',
'size': 5497558138880,
'source': None}Getting information about a specific volume
Now that you have a volume name, you can get information about a specific volume. Use a rest API call from the documentation to get that information. See the following:
In [12]:
session.get('https://slc-coz.slc.purestorage.com/api/1.12/volume/ISO-Repository').text
Out[12]:
'{"source": null, "serial": "4966C24FCE8A51E90001007E", "created": "2014-01-23T00:11:00Z", "name": "ISO-Repository", "size": 5497558138880}'Tracking volume growth over time
You can save these outputs and track volume growth over time.
In [13]:
# Run a query to the rest API every 60 seconds.
volume_over_time = []
for i in range(10):
timestamp = pandas.Timestamp.now()
size = session.get('https://slc-coz.slc.purestorage.com/api/1.12/volume/ISO-Repository').json()['size']
volume_over_time.append({'time': timestamp, 'size': size})
time.sleep(1)
In [14]:
# Convert our list of dictionaries to a dataframe for easy plotting
df = pandas.DataFrame(volume_over_time)
# Set our axes to time for simple plotting.
df.set_index(['time'], inplace=True)
In [15]:
df.head()
Out[15]:
|
size |
|
|---|---|
|
time |
|
|
2018-02-21 11:19:14.594602 |
5497558138880 |
|
2018-02-21 11:19:15.632903 |
5497558138880 |
|
2018-02-21 11:19:16.667310 |
5497558138880 |
|
2018-02-21 11:19:17.689911 |
5497558138880 |
|
2018-02-21 11:19:18.718040 |
5497558138880 |
In [16]:
ax = df.plot(title='ISO-Repository size over time')
ax.set_xlabel('Time')
ax.set_ylabel('Size (bytes)')
Out[16]:
Text(0,0.5,'Size (bytes)')
Getting information about multiple volumes
We could do this for every volume over time instead of individual ones.
In [17]:
# Run a query to the rest API every 60 seconds.
volumes_over_time = []
for i in range(10):
size_data = session.get('https://slc-coz.slc.purestorage.com/api/1.12/volume').json()
timestamp = pandas.Timestamp.now()
for size_datum in size_data:
size_datum['time'] = timestamp
volumes_over_time.extend(size_data)
time.sleep(1)
In [19]:
volumes_over_time[0:6]
Out[19]:
[{'created': '2014-04-18T11:45:12Z',
'name': '(null)',
'serial': '4966C24FCE8A51E90000FFFE',
'size': 322122547200,
'source': None,
'time': Timestamp('2018-02-21 11:21:37.257984')},
{'created': '2014-01-23T00:11:00Z',
'name': 'ISO-Repository',
'serial': '4966C24FCE8A51E90001007E',
'size': 5497558138880,
'source': None,
'time': Timestamp('2018-02-21 11:21:37.257984')},
{'created': '2014-04-12T04:32:24Z',
'name': 'JhopFC1',
'serial': '4966C24FCE8A51E9000100F3',
'size': 4398046511104,
'source': None,
'time': Timestamp('2018-02-21 11:21:37.257984')},
{'created': '2014-04-12T04:32:34Z',
'name': 'JhopFC2',
'serial': '4966C24FCE8A51E9000100F4',
'size': 4398046511104,
'source': None,
'time': Timestamp('2018-02-21 11:21:37.257984')},
{'created': '2014-04-17T20:52:46Z',
'name': 'Jhop_RDM',
'serial': '4966C24FCE8A51E9000100F8',
'size': 1099511627776,
'source': None,
'time': Timestamp('2018-02-21 11:21:37.257984')},
{'created': '2014-11-24T22:11:54Z',
'name': 'ESXi-VDI-LUN10',
'serial': '4966C24FCE8A51E900011210',
'size': 21990232555520,
'source': None,
'time': Timestamp('2018-02-21 11:21:37.257984')}]