No momento, esta página está disponível apenas em inglês.
Developer working on computer

If you’re a data scientist, machine learning engineer, or data engineer, you know the challenges of deploying a successful machine learning (ML) model. Difficulties in testing models and the iterative process of training, testing, and validating a model can be a tedious process that continues long after deployment to production. This can mean serious consequences for your team and business, as failure to detect and address low performance can mean the difference between a model that delivers exceptional performance over time and one that fails to deliver business value.

That’s why New Relic is partnering with Comet to bring you an integration that extends full-stack observability to machine learning models and allows you to establish production performance baselines based on model performance.

In the following short demo, you’ll see how you can integrate machine learning metrics with New Relic’s integration with Comet to continually monitor your data across the full machine learning lifecycle. This helps ensure optimal model performance and can help you achieve better business outcomes.

Why should you integrate Comet with New Relic One?

Comet is an MLOps platform that enables teams to automatically monitor machine learning experiments and models in production with just a few lines of code. 

New Relic-Comet integration

Comet’s integration with New Relic One extends infrastructure observability to include your ML models in production and their performance baselines established in training, so you can build better models faster.

The New Relic integration with Comet empowers you to:

  1. Observe: Retrain your models faster by observing your models from production to deployment. 
  2. Debug: Use powerful code-based metrics and visualizations to debug faster.
  3. Review: Detect issues and address performance degradation, data drift, and potential bias by reviewing your ML metrics.
  4. Validate: Compare models in production against their baselines during training to validate performance.
  5. Adapt: React to changes in performance caused by model drift.
  6. Correlate: Correlate ML production issues with the rest of your software and infrastructure stack.
  7. Collaborate: Work across teams seamlessly by breaking down the silos between data scientists and developers to ensure the accuracy of models in production, all in one place.

You’ll get data and insights to help build better, more accurate ML models while also improving productivity, collaboration, and visibility across your team.

Integrating Comet with New Relic One

To set up the Comet integration, you need a New Relic One account. If you don’t already have one, sign up for a forever free account.

1. Log in to, go to the Explorer page, and select + Add more data.

2. Select Comet from the MLOps Integration section.

Comet integration button on dashboard

3. Select your New Relic Account ID

Select your account id in the UI dropdown

4. Create or select an existing Insight API key from the Real Time Training Metric section. 

Create an API key

5. Contact comet at to set up the integration and view their MPM Dashboard. You will need the Insight API Key.  

Comet's dashboard

6. View the Comet dashboard in New Relic to start tracking the performance of your machine learning models in New Relic.