Real-World MLOps for Batch Inference with Model Monitoring Using Open Source Technologies

May 3, 2024

In this tutorial, we will set up monitoring and metrics collection for an AI model endpoint using open-source technologies Prometheus and Grafana. The endpoint will be built with Flask and Ollama, providing a simple API for prompting an AI assistant. 

The Need to Monitor a Model Endpoint

Monitoring a model endpoint is crucial in MLOps for several reasons:

1. Performance Tracking: This is done to ensure that the model is performing as expected post-deployment. Performance can degrade over time due to concept drift, data drift, or changes in the environment where the model is deployed.

2. Data Quality: We need to monitor the model for data quality and check if there are anomalies in input data. Anomalies might indicate issues with data pipelines or unexpected changes in data patterns if the model wasn't trained to handle them.

3. Resource Utilization: To monitor the compute resources being used by the model endpoint, such as CPU, memory, and disk I/O, this helps in optimizing resource allocation and can reduce costs.

4. Latency: Monitoring helps in measuring the response time of the model endpoint. High latency could affect user experience and the overall performance of the application using the model.

5. Throughput: Model monitoring helps in tracking the number of requests that the model endpoint can handle in a given time frame. This helps in understanding the scalability needs and whether the current infrastructure is sufficient.

6. Error Rates: It is also used to identify the rate of failed requests or prediction errors, which could indicate problems with the model or with the infrastructure.

7. Versioning: It manages and keeps track of different versions of models that are deployed, making it easier to roll back to previous versions if needed.

8. Security: It ensures that the model endpoint is secure from external threats and that data privacy is maintained, especially if sensitive data is being used.

9. Compliance and Auditing: It maintains logs and records of the model's predictions and performance, which may be necessary for regulatory compliance in certain industries.

10. Feedback Loop: It collects information on the model's performance that can be used to further refine and improve the model through retraining.

In essence, monitoring a model endpoint is about maintaining the quality and reliability of machine learning services in production. It enables the team to respond quickly to issues, update models as needed, and ensure that the model meets its objectives.

In this article, we will use the open-source technologies Prometheus and Grafana to monitor a Llama2 endpoint deployed using Ollama and served via the Flask framework.

A Note on Prometheus and Grafana 

Prometheus is an open-source monitoring and alerting toolkit widely used in dynamic service-oriented environments. It was originally built by SoundCloud and has a large community of developers and users. Prometheus is designed to collect and process metrics in real time. Prometheus stores data as time series, and each time series is identified by a metric name and a set of key-value pairs, known as labels.

Grafana is an open-source analytics and interactive visualization web application. It provides charts, graphs, and alerts for the web when connected to supported data sources, one of the most popular being Prometheus. Grafana allows users to create insightful and beautiful dashboards that are customizable and shareable.

Prometheus and Grafana are often used together to deliver effective monitoring and visualization capabilities. Prometheus collects and stores metrics, while Grafana provides a powerful interface to visualize the data stored in Prometheus. The combination allows developers and system administrators to detect and respond to issues in their environments, understand system behavior, and improve the performance of applications. Grafana's ability to pull data from multiple sources also allows for a combined view across different systems and components, providing a comprehensive overview of the infrastructure's health and performance.

Launching a GPU Node 

Since we will be working with a local model pulled from Ollama, we will be needing a GPU with sufficiently large memory. For running the code in this blog, I used a V100 GPU node launched on E2E networks

Head over to to sign up.

Getting into the Code

Start an Ollama Server.

OLLAMA_HOST= ollama serve

This will start an instance of Ollama on the default port 11434.

Then pull the llama2:7b model.

curl http://localhost:11434/api/pull -d '{
  "name": "llama2"

Ensure the following dependencies.

pip install flask ollama prometheus_client flask_prometheus_metrics

Now create a file called and paste the following into it.

from flask import Flask, request, Response, stream_with_context
from prometheus_client import make_wsgi_app
from werkzeug.middleware.dispatcher import DispatcherMiddleware
from werkzeug.serving import run_simple
from flask_prometheus_metrics import register_metrics
from ollama import Client

app = Flask(__name__)

# Initialize the Ollama Client outside of the request to avoid reinitialization on every request
client = Client(host='http://localhost:11434')

@app.route('/generate', methods=['POST'])
def generate():
    # Get the prompt from the request's data
    data = request.json
    prompt = data.get('prompt')

    if not prompt:
        return Response("No prompt provided", status=400)

    # Streaming the response using Ollama
    def generate_stream(prompt):
            for token in'llama2:7b', messages=[{'role': 'user', 'content': prompt}], stream=True):
                # Use yield to stream the response
                yield token['message']['content']
        except Exception as e:
            # Stop the stream and handle the exception if necessary
            yield str(e)

    # The stream_with_context decorator is necessary to keep the context around the generator alive
    return Response(stream_with_context(generate_stream(prompt)), content_type='text/plain')

# provide app's version and deploy environment/config name to set a gauge metric
register_metrics(app, app_version="v0.1.2", app_config="staging")

# Plug metrics WSGI app to your main app with dispatcher
dispatcher = DispatcherMiddleware(app.wsgi_app, {"/metrics": make_wsgi_app()})

run_simple(hostname="", port=5000, application=dispatcher)

The given Python code is for a Flask web application that integrates with Prometheus for metrics monitoring while using Ollama to generate responses to prompts. The code is broken down into several parts:

1. Imports: The necessary modules and functions are imported from Flask, Prometheus, Werkzeug, a custom Flask Prometheus metrics utility, and `ollama`.

2. Flask App Initialization: A Flask app instance is created.

3. Ollama Client Initialization: A client for the Ollama service is initialized outside of the request handling function to prevent reinitialization for every request, which would be inefficient.

4. Route Definition (`/generate`):

   - The `/generate` endpoint is defined to accept POST requests. 

   - It extracts a `prompt` from the JSON body of the request.

   - If no prompt is provided, it returns a 400 Bad Request response.

   - If a prompt is provided, it calls Ollama to generate content based on the prompt, streaming the response. This is particularly useful for long-running generation tasks, as it can provide output incrementally.

   - An inner function `generate_stream` is defined to use the Ollama client to generate responses. It uses a generator and the `yield` keyword to stream the response back to the client.

   - If an exception occurs during generation, it is caught, and the exception message is sent back in the stream.

   - The Flask `stream_with_context` decorator is used to ensure that the Flask context is not lost during streaming. This decorator is necessary because the streaming process is a generator, and without the decorator, Flask might lose the context of the current application or request.

   - The `/generate` endpoint returns a `Response` object with streaming content.

5. Prometheus Metrics Registration:

   - The `register_metrics` function is called, adding Prometheus metrics tracking to the Flask app. The application version (`v0.1.2`) and configuration (`staging`) are passed as labels for these metrics.

6. DispatcherMiddleware:

   - The `DispatcherMiddleware` is used to combine the Flask app with another WSGI application. Here, it is set up to route requests to `/metrics` to a Prometheus WSGI application created by `make_wsgi_app()`. This allows the Flask app to serve its endpoints while also serving the Prometheus metrics endpoint under the same WSGI server.

7. Running the Application:

   - The `run_simple` function from Werkzeug is used to run a development WSGI server. The server listens on all available IP addresses (``) at port `5000`, serving the combined Flask and Prometheus app via the `dispatcher`.

This application setup allows for a simple AI content generation service with streaming responses and monitoring capabilities via Prometheus. The Flask app serves the application's main functionality and the metrics endpoint, making it a self-contained service suitable for development and possibly staging environments.

Now run the Flask application.


A Flask server will be up and running on http://localhost:5000. You can query your llama2:7b model by sending a cURL request as follows:

curl -X POST http://localhost:5000/generate     -H "Content-Type: application/json"     -d '{"prompt": "Write me a long poem!"}'

You will see the tokens streaming in through your terminal window as below.

Now create a file called docker-compose.yaml and paste the following into it.

ersion: "3"
    image: prom/prometheus:latest
      - "9090:9090"
      - "./prometheus.yml:/etc/prometheus/prometheus.yml"

The `docker-compose.yml` file defines how your Prometheus service should run within a Docker container, including which image to use, which ports to expose, and where to find its configuration file.

Make another file called prometheus.yml and paste the following into it.

global:  scrape_interval: 15s  evaluation_interval: 15sscrape_configs:  - job_name: 'prometheus'    static_configs:      - targets: ['']‍

The `prometheus.yml` file is the configuration for Prometheus itself, detailing how it should scrape metrics. In the targets, we need to plug in the <server-ip>:<port> of our flask application since Prometheus will try to scrape it externally. 

Now run the following commands:

sudo docker-compose buildsudo docker-compose up

This will start a Prometheus server at http://<server-ip>:9090

You can go to the target endpoint of this server to see that it is monitoring your Flask application.

Latest Blogs
This is a decorative image for: A Complete Guide To Customer Acquisition For Startups
October 18, 2022

A Complete Guide To Customer Acquisition For Startups

Any business is enlivened by its customers. Therefore, a strategy to constantly bring in new clients is an ongoing requirement. In this regard, having a proper customer acquisition strategy can be of great importance.

So, if you are just starting your business, or planning to expand it, read on to learn more about this concept.

The problem with customer acquisition

As an organization, when working in a diverse and competitive market like India, you need to have a well-defined customer acquisition strategy to attain success. However, this is where most startups struggle. Now, you may have a great product or service, but if you are not in the right place targeting the right demographic, you are not likely to get the results you want.

To resolve this, typically, companies invest, but if that is not channelized properly, it will be futile.

So, the best way out of this dilemma is to have a clear customer acquisition strategy in place.

How can you create the ideal customer acquisition strategy for your business?

  • Define what your goals are

You need to define your goals so that you can meet the revenue expectations you have for the current fiscal year. You need to find a value for the metrics –

  • MRR – Monthly recurring revenue, which tells you all the income that can be generated from all your income channels.
  • CLV – Customer lifetime value tells you how much a customer is willing to spend on your business during your mutual relationship duration.  
  • CAC – Customer acquisition costs, which tells how much your organization needs to spend to acquire customers constantly.
  • Churn rate – It tells you the rate at which customers stop doing business.

All these metrics tell you how well you will be able to grow your business and revenue.

  • Identify your ideal customers

You need to understand who your current customers are and who your target customers are. Once you are aware of your customer base, you can focus your energies in that direction and get the maximum sale of your products or services. You can also understand what your customers require through various analytics and markers and address them to leverage your products/services towards them.

  • Choose your channels for customer acquisition

How will you acquire customers who will eventually tell at what scale and at what rate you need to expand your business? You could market and sell your products on social media channels like Instagram, Facebook and YouTube, or invest in paid marketing like Google Ads. You need to develop a unique strategy for each of these channels. 

  • Communicate with your customers

If you know exactly what your customers have in mind, then you will be able to develop your customer strategy with a clear perspective in mind. You can do it through surveys or customer opinion forms, email contact forms, blog posts and social media posts. After that, you just need to measure the analytics, clearly understand the insights, and improve your strategy accordingly.

Combining these strategies with your long-term business plan will bring results. However, there will be challenges on the way, where you need to adapt as per the requirements to make the most of it. At the same time, introducing new technologies like AI and ML can also solve such issues easily. To learn more about the use of AI and ML and how they are transforming businesses, keep referring to the blog section of E2E Networks.

Reference Links

This is a decorative image for: Constructing 3D objects through Deep Learning
October 18, 2022

Image-based 3D Object Reconstruction State-of-the-Art and trends in the Deep Learning Era

3D reconstruction is one of the most complex issues of deep learning systems. There have been multiple types of research in this field, and almost everything has been tried on it — computer vision, computer graphics and machine learning, but to no avail. However, that has resulted in CNN or convolutional neural networks foraying into this field, which has yielded some success.

The Main Objective of the 3D Object Reconstruction

Developing this deep learning technology aims to infer the shape of 3D objects from 2D images. So, to conduct the experiment, you need the following:

  • Highly calibrated cameras that take a photograph of the image from various angles.
  • Large training datasets can predict the geometry of the object whose 3D image reconstruction needs to be done. These datasets can be collected from a database of images, or they can be collected and sampled from a video.

By using the apparatus and datasets, you will be able to proceed with the 3D reconstruction from 2D datasets.

State-of-the-art Technology Used by the Datasets for the Reconstruction of 3D Objects

The technology used for this purpose needs to stick to the following parameters:

  • Input

Training with the help of one or multiple RGB images, where the segmentation of the 3D ground truth needs to be done. It could be one image, multiple images or even a video stream.

The testing will also be done on the same parameters, which will also help to create a uniform, cluttered background, or both.

  • Output

The volumetric output will be done in both high and low resolution, and the surface output will be generated through parameterisation, template deformation and point cloud. Moreover, the direct and intermediate outputs will be calculated this way.

  • Network architecture used

The architecture used in training is 3D-VAE-GAN, which has an encoder and a decoder, with TL-Net and conditional GAN. At the same time, the testing architecture is 3D-VAE, which has an encoder and a decoder.

  • Training used

The degree of supervision used in 2D vs 3D supervision, weak supervision along with loss functions have to be included in this system. The training procedure is adversarial training with joint 2D and 3D embeddings. Also, the network architecture is extremely important for the speed and processing quality of the output images.

  • Practical applications and use cases

Volumetric representations and surface representations can do the reconstruction. Powerful computer systems need to be used for reconstruction.

Given below are some of the places where 3D Object Reconstruction Deep Learning Systems are used:

  • 3D reconstruction technology can be used in the Police Department for drawing the faces of criminals whose images have been procured from a crime site where their faces are not completely revealed.
  • It can be used for re-modelling ruins at ancient architectural sites. The rubble or the debris stubs of structures can be used to recreate the entire building structure and get an idea of how it looked in the past.
  • They can be used in plastic surgery where the organs, face, limbs or any other portion of the body has been damaged and needs to be rebuilt.
  • It can be used in airport security, where concealed shapes can be used for guessing whether a person is armed or is carrying explosives or not.
  • It can also help in completing DNA sequences.

So, if you are planning to implement this technology, then you can rent the required infrastructure from E2E Networks and avoid investing in it. And if you plan to learn more about such topics, then keep a tab on the blog section of the website

Reference Links

This is a decorative image for: Comprehensive Guide to Deep Q-Learning for Data Science Enthusiasts
October 18, 2022

A Comprehensive Guide To Deep Q-Learning For Data Science Enthusiasts

For all data science enthusiasts who would love to dig deep, we have composed a write-up about Q-Learning specifically for you all. Deep Q-Learning and Reinforcement learning (RL) are extremely popular these days. These two data science methodologies use Python libraries like TensorFlow 2 and openAI’s Gym environment.

So, read on to know more.

What is Deep Q-Learning?

Deep Q-Learning utilizes the principles of Q-learning, but instead of using the Q-table, it uses the neural network. The algorithm of deep Q-Learning uses the states as input and the optimal Q-value of every action possible as the output. The agent gathers and stores all the previous experiences in the memory of the trained tuple in the following order:

State> Next state> Action> Reward

The neural network training stability increases using a random batch of previous data by using the experience replay. Experience replay also means the previous experiences stocking, and the target network uses it for training and calculation of the Q-network and the predicted Q-Value. This neural network uses openAI Gym, which is provided by taxi-v3 environments.

Now, any understanding of Deep Q-Learning   is incomplete without talking about Reinforcement Learning.

What is Reinforcement Learning?

Reinforcement is a subsection of ML. This part of ML is related to the action in which an environmental agent participates in a reward-based system and uses Reinforcement Learning to maximize the rewards. Reinforcement Learning is a different technique from unsupervised learning or supervised learning because it does not require a supervised input/output pair. The number of corrections is also less, so it is a highly efficient technique.

Now, the understanding of reinforcement learning is incomplete without knowing about Markov Decision Process (MDP). MDP is involved with each state that has been presented in the results of the environment, derived from the state previously there. The information which composes both states is gathered and transferred to the decision process. The task of the chosen agent is to maximize the awards. The MDP optimizes the actions and helps construct the optimal policy.

For developing the MDP, you need to follow the Q-Learning Algorithm, which is an extremely important part of data science and machine learning.

What is Q-Learning Algorithm?

The process of Q-Learning is important for understanding the data from scratch. It involves defining the parameters, choosing the actions from the current state and also choosing the actions from the previous state and then developing a Q-table for maximizing the results or output rewards.

The 4 steps that are involved in Q-Learning:

  1. Initializing parameters – The RL (reinforcement learning) model learns the set of actions that the agent requires in the state, environment and time.
  2. Identifying current state – The model stores the prior records for optimal action definition for maximizing the results. For acting in the present state, the state needs to be identified and perform an action combination for it.
  3. Choosing the optimal action set and gaining the relevant experience – A Q-table is generated from the data with a set of specific states and actions, and the weight of this data is calculated for updating the Q-Table to the following step.
  4. Updating Q-table rewards and next state determination – After the relevant experience is gained and agents start getting environmental records. The reward amplitude helps to present the subsequent step.  

In case the Q-table size is huge, then the generation of the model is a time-consuming process. This situation requires Deep Q-learning.

Hopefully, this write-up has provided an outline of Deep Q-Learning and its related concepts. If you wish to learn more about such topics, then keep a tab on the blog section of the E2E Networks website.

Reference Links

This is a decorative image for: GAUDI: A Neural Architect for Immersive 3D Scene Generation
October 13, 2022

GAUDI: A Neural Architect for Immersive 3D Scene Generation

The evolution of artificial intelligence in the past decade has been staggering, and now the focus is shifting towards AI and ML systems to understand and generate 3D spaces. As a result, there has been extensive research on manipulating 3D generative models. In this regard, Apple’s AI and ML scientists have developed GAUDI, a method specifically for this job.

An introduction to GAUDI

The GAUDI 3D immersive technique founders named it after the famous architect Antoni Gaudi. This AI model takes the help of a camera pose decoder, which enables it to guess the possible camera angles of a scene. Hence, the decoder then makes it possible to predict the 3D canvas from almost every angle.

What does GAUDI do?

GAUDI can perform multiple functions –

  • The extensions of these generative models have a tremendous effect on ML and computer vision. Pragmatically, such models are highly useful. They are applied in model-based reinforcement learning and planning world models, SLAM is s, or 3D content creation.
  • Generative modelling for 3D objects has been used for generating scenes using graf, pigan, and gsn, which incorporate a GAN (Generative Adversarial Network). The generator codes radiance fields exclusively. Using the 3D space in the scene along with the camera pose generates the 3D image from that point. This point has a density scalar and RGB value for that specific point in 3D space. This can be done from a 2D camera view. It does this by imposing 3D datasets on those 2D shots. It isolates various objects and scenes and combines them to render a new scene altogether.
  • GAUDI also removes GANs pathologies like mode collapse and improved GAN.
  • GAUDI also uses this to train data on a canonical coordinate system. You can compare it by looking at the trajectory of the scenes.

How is GAUDI applied to the content?

The steps of application for GAUDI have been given below:

  • Each trajectory is created, which consists of a sequence of posed images (These images are from a 3D scene) encoded into a latent representation. This representation which has a radiance field or what we refer to as the 3D scene and the camera path is created in a disentangled way. The results are interpreted as free parameters. The problem is optimized by and formulation of a reconstruction objective.
  • This simple training process is then scaled to trajectories, thousands of them creating a large number of views. The model samples the radiance fields totally from the previous distribution that the model has learned.
  • The scenes are thus synthesized by interpolation within the hidden space.
  • The scaling of 3D scenes generates many scenes that contain thousands of images. During training, there is no issue related to canonical orientation or mode collapse.
  • A novel de-noising optimization technique is used to find hidden representations that collaborate in modelling the camera poses and the radiance field to create multiple datasets with state-of-the-art performance in generating 3D scenes by building a setup that uses images and text.

To conclude, GAUDI has more capabilities and can also be used for sampling various images and video datasets. Furthermore, this will make a foray into AR (augmented reality) and VR (virtual reality). With GAUDI in hand, the sky is only the limit in the field of media creation. So, if you enjoy reading about the latest development in the field of AI and ML, then keep a tab on the blog section of the E2E Networks website.

Reference Links

Build on the most powerful infrastructure cloud

A vector illustration of a tech city using latest cloud technologies & infrastructure