[Proposal] - Contextual acceleration

[shanilpa]

Overview

The following proposal aims to show an experience where users are able to accelerate external data for better performance in Discover.

There are two flows to consider:

1. Acceleration in the same Discover tab: These are mainly to support use cases for analysts that do not have technical knowledge on setting up acceleration from the ground up. These flows pre-populate the majority of acceleration fields and only require reviewing or light configuration.
2. Acceleration in a new browser tab: These are mainly to support use cases where not enough information can be pre-populated and requires advanced configurations or a user is running a process in Discover (i.e query) but might want to setup acceleration without cancelling that process.

Acceleration Flyout

Users can accelerate external data in Discover via a flyout. This flyout is designed for analysts and simplifies acceleration flows to make it easier for mid-high knowledge (about accelerations) users to accelerate data if they have the permissions to do so.

Expanded view

Covering index

Materialized view

Acceleration from data selection

Acceleration in a new browser tab

Users are able to access the acceleration flow from Discover via the data selection modal. This would open a new tab with the acceleration flow with items that can be pre-populated populated in the flow.

Acceleration on loading states

Acceleration in a new browser tab

For queries to external data sources that are taking time to execute (over 10s) users are prompted with information on acceleration and a pathway to open an acceleration flow in a new tab with items that can be pre-populated populated. They are also presented with a link to learn more about acceleration.

Acceleration from the inspect panel

Acceleration in same Discover tab

For users that executed a query and are inspecting it’s performance they are presented with an option in the inspect flyout to accelerate their data for better performance.

Acceleration for visualizations

For a user to create visualizations in Discover they have to have an OpenSearch index as the data source (covering index or materialized view). If users do not have these and they select the visualize tab we notify them and offer a path to creating accelerations in a new tab. Information that can be pre-populated will be populated in the new tab.

[kgcreative]

cc: @anirudha