What’s New: Dataset Graph

By Knox Lively, November 8, 2021

Since our most recent product update announcement in August, we’ve squashed hundreds of bugs and added dozens of new features and improvements. Though we continuously release updates and new features to our product, we recently released a feature set we simply can’t stay quiet about. 

In the past week, we released a group of features we call Dataset Graph. This new group of features allows users to visually, and easily, see relationships between datasets using a user-friendly graph. Each new feature found in Dataset Graph essentially offers the same functionality, but with drastically different ways to visualize the same data.

Dataset Graph is an extension of GraphLink, which we currently use in Observe to provide one-click navigation from one related dataset to another. Users find this feature drastically reduces investigation time by providing them with immediate access to relevant and contextual information. As useful as GraphLink is, we knew there was still room to improve. 

Where is Dataset Graph?

Dataset Graph can easily be found in the navigation rail to the left in your Observe workspace under the “Explore” tab. From there you will find a new tab at the top of the page, unsurprisingly, called “Dataset Graph”.

Dataset Graph location in Observe SaaS Platform


As the name implies, the Links view in Dataset Graph allows you to quickly and elegantly see links between datasets. 

Dataset Graph Explore View Links in Observe SaaS Product

In this particular view, all datasets are first shown as a cluster configuration, grouping linked datasets more closely together than more distantly related ones. Upon clicking a dataset, you get an enhanced view, where relationship links between datasets are highlighted.


The Lineage view offers a different way to view your datasets and their relationships as a flowchart. This view helps offer visibility into how datasets are constructed (their “parents”), as well as how they inform other datasets (their “children”).

Dataset Graph Explore View Lineage in Observe SaaS Product

This view is particularly useful when constructing new datasets, or tracking down the source of an alert. It may also be useful to think of Lineage as a genealogy map for your data.


Much like the other views in Dataset Graph, Focus also shows you relationships between your datasets. This particular view allows you to inspect a single dataset to more clearly see direct links between it and other datasets, however, this particular view allows you to inspect a singular dataset to more clearly see direct links between it and other datasets.

Dataset Graph Explore View with Focus in Observe SaaS Product

Tabbed Dataset Graph

In addition to accessing Dataset Graph via the Explore page, we’ve also made it conveniently accessible directly from a dataset or a resource via the Related tab. This is particularly useful when performing data modeling, as there is no need to navigate to a separate page to see relationships to the current dataset.

Dataset Graph Tabbed View in Observe SaaS Product


Lastly, you can see the “status” of datasets with a simple overlay by selecting the “Show Status” box. With this, you can quickly identify broken relationships between datasets, drastically reducing the time it takes to resolve issues while creating new datasets or hunting down alerts.

Dataset Graph Explore View Show Statusin Observe SaaS Product

Wrapping Up

We believe that the new feature set will make users more efficient than ever, as they can now easily see which datasets are connected — or not. This empowers SRE’s, DevOps Engineers, or whomever to make the most of the data they already have by linking existing datasets together to drive new insights. More importantly, engineers can add this value without adding more tools, and subsequently, more cost to their already complicated environment(s). 

We’re excited to see what our users will be able to do as a result of adding Dataset Graph to our already powerful platform. If you’d like to know more about Observe, we encourage you to check out our demo.