All Articles

Client story: PacketAI

Founded in 2019 and having already raised over 2.1M Euros PacketAI is on its way to take the IT monitoring industry to the next level as the first ROI driven ITOps solution. PacketAI Using advanced machine learning algorithms it helps predict IT incidents in your infrastructure and gets to the root causes of the incidents that have already happened. It connects to your systems using numerous plug and play integrations and starts parsing your data through powerful machine learning algorithms enabling you to monitor your cloud services, on-premise servers, databases and more from one place through an intuitive interface.

Monitoring is hard and determining the relevance of received data is a monumental task when dealing with big data. PacketAI effectively removes the noise and lets your engineers only deal with data that is relevant saving you precious time in handling unwanted events throughout your infrastructure. It then analyses your data and leverages the power of deep learning to predict incidents before they actually impact your customers. Such predictions end up in a massive reduction in average repair time by up to 50%.

Challenges and solution

Tilde Loop has been chosen as a partner in building and designing parts of PacketAI’s IT monitoring solution. The requirements for building such a large system encompass a plethora of technologies utilizing the most modern and reliable tools. From Kafka to handle a neverending stream of message logs to elasticsearch which enable users realtime search over huge amounts of data. The engineering challenges come with saving and processing huge amounts of data, making the service reliable under heavy loads while also keeping it highly extensible and easily maintainable making Node.js a natural selection to deal with the ever-increasing requirements.

PacketAI’s goal is to provide intuitive integrations with most of the popular services like AWS, GitLab, Kubernetes, Docker or anything that can output logs by enabling the user data collection and analysis at a click of a button.

These requirements coupled with the complexity of the data provide a challenge in building a user interface that is both filled with functionalities and easy is to grasp for a regular engineer who can manipulate it in a seamless manner while getting the most succinct and effective result.

Building such interfaces requires a steady influx of usability feedback with frequent changes in the UI until it gets to a sweet spot where data meets visualization at the benefit of the customer.

This is achieved using React by creating a clean front-end interface without taking away from the details and taking into consideration the growth of data as well as supported implementations. Having a clean feature-rich dashboard coupled with visualizations using numerous graph libraries such as D3.js makes PacketAI a natural expansion of the workplace of a typical engineer.

The solution, of course, needs to be packed into a clean, expandable and maintainable codebase based on the top industry standards.

We are working closely to achieve all of these standards and are committed to finding solutions for future features and challenges that are ahead of PacketAI, a fresh face in the IT monitoring industry.

Like what you've read? Join our newsletter

Launching a new JavaScript project? Need help on an existing project? Work with us