[CASSSIDECAR-174] Mechanism to have a reduced number of Sidecar instances run operations

• Type: Improvement
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 1.0
• Component/s: Rest API
• Change Category: Semantic
• Complexity: Normal

Source Control Link

https://github.com/apache/cassandra-sidecar/commit/ba7e58c43c334260f296a28079ad3a1d5a3e3605

Test and Documentation Plan

Add integration tests covering the new code and unit tests covering the additional configuration code.

• Assignee: Francisco Guerrero
• Reporter: Francisco Guerrero

Authors

Francisco Guerrero

Reviewers

Doug Rohrer, Yifan Cai

• Created: 10/Dec/24 23:30
• Updated: 19/Dec/24 23:56
• Resolved: 19/Dec/24 23:56

Description

This is a proposal to introduce a mechanism to have a reduced number of Sidecar instances run operations across the cluster. Ideally we would want a unique Sidecar instance perform a subset of operations in the cluster. However, the operations might be triggered by more than one Sidecar instance so they can be assumed to be safe to be performed by one or more leaders.
 
Currently Sidecar has existing use cases where we want a single Sidecar instance to perform, for example schema initialization. There are upcoming use cases in Sidecar that might be able to take advantage of this feature.

Goals

• Have the ability to choose a reduced number of instances that can perform operations where we would want only a minimal number of instances perform the operations
• Easily write new tasks or define schemas that are going to be executed by the reduced number of Sidecar instances
• Have some loose guarantees on always having at least one leader

Non-Goals

• Have a unique global leader

Issue Links

• links to: GitHub Pull Request #160