Using Postman to create multiple Websites via API & CSV?

See, I don’t know Python.

Learn how to do API calls in PS vs. learn how to do Python. 6’s really. I get it though. I’m having to learn two new tools and mongodb. I can’t count the number of times I’ve said, “see, if it just used jinja2 templating, i’d be done by now”.

I’m pushing Python so I don’t have to learn Powershell just to be able to help you. ;-)

I’d take a Python shell.  I guess I could just use the one from the LM example page, but those are old.

All I’d have to do is change the APIKey stuff and then sub in my “Data” information, right?  One of the main reasons I was learning Python is for API stuff, but I’m just going through an online course so I don’t know when we’ll get to that part.  ;)

I’d take a Python shell.  I guess I could just use the one from the LM example page, but those are old.

Yeah, it should work. If it doesn’t maybe they’ll finally update them and/or start showing examples using the SDK. Make sure to add the header to force it to v3 of the API:

#!/bin/env python

import requests
import json
import hashlib
import base64
import time
import hmac

#Account Info
AccessId ='API-ACCESS-ID'
AccessKey ='API-ACCESS-KEY'
Company = 'apiAccount'

#Request Info: Add a service
httpVerb ='POST'
resourcePath = '/service/services'
data = '{"type":"ping","name":"LMCertification","disableAlerting":false,"pollingInterval":5,"host":"www.logicmonitor.docebosaas.com"}'

#Construct URL 
url = 'https://'+ Company +'.logicmonitor.com/santaba/rest' + resourcePath

#Get current time in milliseconds
epoch = str(int(time.time() * 1000))

#Concatenate Request details
requestVars = httpVerb + epoch + data + resourcePath

#Construct signature
signature = base64.b64encode(hmac.new(AccessKey,msg=requestVars,digestmod=hashlib.sha256).hexdigest())

#Construct headers
auth = 'LMv1 ' + AccessId + ':' + signature + ':' + epoch
headers = {'Content-Type':'application/json','Authorization':auth,'X-Version':'3'}

#Make request
response = requests.post(url, data=data, headers=headers)

#Print status and body of response
print 'Response Status:',response.status_code
print 'Response Body:',response.content

Suggest you not use the example destination. That website hasn’t existed for years (at least 4 that I know of).

You can, of course, take a shell from our support docs, but unfortunately our docs haven’t quite kept pace with the development of our API, and most of those scripts reflect v1 or v2. The script from the repo I linked yesterday (https://github.com/krshearman/LMAPI_v3/blob/main/PostAddWebsite.py) has been recently tested and is working.  All you’ll need to do is add your company name, creds, and your own data. 

Hope it helps@Kelemvor !

I disagree that people new to Python should first learn the Requests method of interacting with LM’s API. That’s like making people learn how to rebuild an internal combustion engine before being allowed to buy an electric car.

Use the SDK; it’s object oriented and much easier to get started with, especially for new Python programmers.

I had a similar experience when learning Python. I was trying to learn how to code robots for the Raspberry Pi. I took all their examples, learned all the long methods and studied how those methods perfectly aligned with the physical hardware in a very bare metal way. It was hard, but I was able to do what I needed to do. Then the gpiozero library came out, which is an SDK for the bare metal GPIO calls. Some of my code blocks went from 50-70 lines to two (2). Every time I built something new using gpiozero, I was astounded at how much more simple things were. I didn’t have to create a pulldown resistor object, then instantiate another object with that object and the pin object just so i could call it. I could just create an LED and pass it the pin number and voila, I had a working LED.

Just because the Requests method of interacting with the API exists doesn’t mean it’s the right way for everyone or even most people. Most people who are using LM aren’t full time developers. They’re infrastructure admins who don’t necessarily want to shift to being developers. They’re part time developers for whom the SDK is a great tool.

 Consider the example I showed earlier, which is a complete working example of adding a website to LM:

from lm import lm

body = {
    "testLocation": {"all": False,"smgIds": [2, 3, 4]}, 
    "groupId": 1, 
    "description": "", 
    "type": "webcheck", 
    "isInternal": False, 
    "name": "google.com", 
    "domain": "google.com", 
    "schema": "https", 
    "steps": [{}], 
    "useDefaultAlertSetting": True
}
response = lm.add_website(body)
print(response)

This makes use of my lmwrapper library to simplify all your API interactions. The `from lm import lm` line imports the sdk, sets up the object, passes in your credentials (from a file you have stored outside the script because you should never put your credentials in the script), and otherwise gets everything ready to go. Body creates the definition you want to add and the lm.add_website() method adds it. This is the entirety of the script needed to add a website using the SDK.

Compare this to any script (even any of mine) that use the request library to interact with the API directly. Not only is the code longer, but there are steps in there that aren’t always understood and should never be coded more than once (that’s what imports are for). 

Shorter code through the SDK is easier to read and understand what is going on.

And I respect your opinion. However, some customers just want to see how the api works or test basic functionality. Also, the requests library is widely used and still actively developed. It’s just an easier way to get started.

Well, I’m not sure what changed, but I was able to get the POST to work and create a website.  However, why would I have success POSTing to a website, but then get an Authentication Error when I change it to a GET to grab website Groups, without changing anything else?

Argh!

I can do a GET just fine with resource groups, alerts, etc, but as soon as I try to do website groups It fails.  Is there something special about that one?

Because the GET does not need the payload, so your signature also shouldn’t include the payload.

@Kelemvor ~ check the differences noted in these screenshots:

So basically, with a get, the query parameters are sent by appending to the url (first screenshot). With a post, the data is sent with the request (second screenshot).