Python SDK - JSON

It’s because the `alerts` list, contains a custom object of type “logicmonitor_sdk.models.alert.Alert”. It’s not a dictionary or list or other python native type that json.JSONEncoder().encode() expects. You can verify this by doing this:

print(alerts[0].__class__)

So, you have to convert each item in the list to a dictionary before you can pass it to the encoder. List comprehension comes in pretty handy here. You can split it out if you want or just do it like this:

# create an instance of the API class
api_instance = logicmonitor_sdk.LMApi(logicmonitor_sdk.ApiClient(configuration))   
id = 138 # Integer |      
size = 1000 # Integer |  (optional)  
offset = 0 # Integer |  (optional)  
filter = f'severity:3,cleared:true,startEpoch>:{startEpoch},endEpoch<:{endEpoch}' # String |  (optional)

#Loop shiznit
alerts = []
end_found = False

while not end_found:
  current = api_instance.get_alert_list_by_device_group_id(id, size=size, offset=offset, filter=filter).items
  alerts += current
  offset += len(current)
  end_found = len(current) != size

#Print and encode result to json. This allows you to use this variable later in an API Put.
with open('./customer/json/cleared_sev3_alerts.json', 'w') as content_output_json:
    result_encoded = json.JSONEncoder(sort_keys=False, indent =4).encode([x.to_dict() for x in alerts])
    content_output_json.write(str(alerts))
    content_output_json.close()

print(len(alerts))

This takes advantage of the SDK’s .to_dict() method that takes LM SDK custom classes and converts them into simple dictionaries.