Skip to content

Latest commit

 

History

History
43 lines (38 loc) · 1.43 KB

File metadata and controls

43 lines (38 loc) · 1.43 KB

Persistence options - Eventflow

Option A - Sidecar by Pod

Pros

  • Simplicity when unfolding along with each replica.
  • Minimum latency when operating locally.

cons

  • Duplicate resources and status in each pod.
  • Complex synchronization between replicas.

When I would apply

  • Low -scale deployments or development environments.

Option B - Centralized Service

Pros

  • Strong consistency and concurrence control.
  • Eventflow independent scale.

cons

  • Add a network latency.
  • Requires high availability.

When I would apply

  • When multiple replicas are needed with immediate shared.

Option C - Object Storage

Pros

  • Scalability practically unlimited.
  • Reduced operating cost.

cons

  • Eventual consistency.
  • Superior latency for reading/writing operations.

Target State

  • Long -term destination for mass loads and durability.

Comparison

Criteria Option A - Sidecar Option B - Centralized Service Option C - Object Storage
Complexity Baja Average Average
Latency Minimal Moderate High
Consistency Local, not shared Strong Eventual
Scalability Linked to Pods Independent Unlimited
Operation POD management Unique Service Managed Service

Decision

The ** option B - centralized service ** is adopted as an immediate solution because it offers strong consistency and allows you to climb without duplicating the State as in option A.