Eric Sloof

I've successfully connected a Raspberry Pi 5 running TensorFlow to VMware Aria Operations, and I'm excited to share the simplicity of the process. The new Management Pack Builder from VMware transforms the creation of custom management packs into a straightforward task. This intuitive tool, which requires no coding skills, invites us to extend the capabilities of our monitoring systems with ease. Join me as we explore how this setup can revolutionize your operational management, making it more efficient and responsive to your needs.

The VMware Aria Operations Management Pack Builder is a self-contained tool designed to facilitate the development of bespoke management packs for VMware Aria Operations. It offers a user-friendly, code-free approach to importing data from external APIs. This tool allows the creation of new resources or the enhancement of existing VMware and third-party resources by adding new data, establishing relationships, and integrating events.

TensorFlow, renowned for its versatility in machine learning, is particularly effective for object detection projects on Raspberry Pi. This lightweight platform seamlessly integrates with the modest hardware of the Raspberry Pi, making it ideal for real-time object detection tasks. By utilizing a camera module with TensorFlow, users can develop efficient, on-device models capable of identifying and categorizing objects in the camera's field of view.


The Raspberry Pi 5 uses a Python script tailored for object detection, processing images from its camera into structured JSON data. Additionally, it operates a web server that presents a REST API to Aria Operations, enabling the collection and statistical analysis of object detection data processed by TensorFlow.

{"person": 1, "bicycle": 0, "car": 17, "motorcycle": 0, "airplane": 0, "bus": 0, "train": 0, "truck": 1, "boat": 0, "traffic": 0, "fire": 0, "stop": 0, "parking": 0, "bench": 0, "bird": 0, "cat": 0, "dog": 0, "horse": 0, "sheep": 0, "cow": 0, "elephant": 0, "bear": 0, "zebra": 0, "giraffe": 0, "backpack": 0, "umbrella": 0, "handbag": 0, "tie": 0, "suitcase": 0, "frisbee": 0, "skis": 0, "snowboard": 0, "sports": 0, "kite": 0, "baseball": 0, "skateboard": 0, "surfboard": 0, "tennis": 0, "bottle": 0, "wine": 0, "cup": 0, "fork": 0, "knife": 0, "spoon": 0, "bowl": 0, "banana": 0, "apple": 0, "sandwich": 0, "orange": 0, "broccoli": 0, "carrot": 0, "hot": 0, "pizza": 0, "donut": 0, "cake": 0, "chair": 0, "couch": 0, "potted": 0, "bed": 0, "dining": 0, "toilet": 0, "tv": 0, "laptop": 0, "mouse": 0, "remote": 0, "keyboard": 0, "cell": 0, "microwave": 0, "oven": 0, "toaster": 0, "sink": 0, "refrigerator": 0, "book": 0, "clock": 0, "vase": 0, "scissors": 0, "teddy": 0, "hair": 0, "toothbrush": 0}

The package file and the corresponding Python code can be accessed and downloaded from my GitHub repository.  You can find guidance on setting up TensorFlow on a Raspberry Pi in a previous article I have authored.

In this article, I’ll describe how to connect to the REST API of VMware Aria Operations. I’ll also demonstrate how to use the Swagger UI to construct API entries in Postman.

The initial access to the AriaOps API is based on authentication with a username, password and authentication source provided in a JSON body. Don’t try to do this with basic authentication, because it will fail. Simply construct a POST call with a JSON body, and get the ops:token. This token is needed for any subsequent REST calls against AriaOps.

After receiving a 200 (OK) response. The ops:token has successfully been retrieved. It will expire after 6 hours. Any subsequent requests must include this token.

To construct new API calls, you can use either the API-Programming-Operations Guide or the Swagger UI. In the Swagger UI, it is fairly easy to test API calls with the “Try it out” feature.

This also includes a valid ops:token in the test request. You can copy the Curl request from Swagger and paste it into a new request in Postman. This automatically translates the curl request into a valid REST API call, and fills all the parameters and creates a valid JSON body.

I've created some sample requests and exported the Postman collection to GitHub. Any comments or suggestions on X please. 

In this series of articles, I’ll describe how to build a management pack that integrates a Raspberry Pi Pico W with Aria Operations. For the creation of the management pack, I’ll use the VMware Aria Operations Management Pack Builder. This appliance automates the complete process of programming the RP2040 input sensors via a REST API into a PAK file. Communications between Aria Operations and the Pi Pico will be done wireless using the Microdot framework. 

Creating a management pack for Aria Operations involves so many steps that I’ve decided to split this project into multiple blog articles. I’m using SimpleMind as a mind map tool to keep track of all the different components. All the source code will be hosted on GitHub. Please enhance or adjust this code to your own needs by forking or copying the code.

Before I started writing these articles, I’ve built a working prototype. The reason I’ve built this prototype was to see if it's even possible to get everything to work. The prototype was also used during a recent training course delivery of the VMware vRealize Operations: Advanced Use Cases [v8.x].

The beta version of the Python code, including a working PAK file and some screenshots, are already available on GitHub. In upcoming articles, I’ll enhance the code and create new versions. Currently, the roadmap contains an API with authentication, dashboards, events, alerts, etc. Feel free to bring up any suggestions on my X channel or in the comments on GitHub.