The ChaosSearch configuration automatically creates a database connection from SQL Analytics/Superset to your Chaos Index data. If you plan to create Superset charts and dashboards for your SQL queries, you must add datasets.
What is a dataset?
Superset defines a dataset as a physical or virtual representation of a SQL database table.
A physical dataset, for ChaosSearch, is a pointer to a specific Refinery view. The physical dataset thus makes each Refinery view that you want to visualize available in dataset lists for chart or dashboard editing.
A virtual dataset is points to data results gathered from a SQL query based on a set of rules that you specify. You can use JOIN and any other SQL operations to build your virtual dataset. You can build a SQL query to organize your columns and data, and you can save that construction as a virtual dataset for visualizations.
Datasets offer a “thin semantic layer” that allows users to make some schema changes in the data represented in the Superset charts and dashboards. Users can perform actions like hiding unwanted columns, setting columns as temporal, filterable, or dimension-type data, and defining special metrics and calculations that can be used in chart and dashboard definitions. These changes are stored as a layer in Superset and do not change the underlying view within ChaosSearch.
For each dataset, you can use the Actions menu to perform the following tasks:
Delete a dataset definition (which does not change or affect the Refinery view)
Export the dataset to save its definition as a YAML file in your local downloads area
Edit the dataset to take advantage of Superset features for customizing column properties, and using the Superset schematic layer for creating custom metrics
Updated about 1 month ago