Skip to content

feat: add connection-pool bug scenario #3

Description

@jaeyeopme

Under sustained load, misconfigured connection pools silently exhaust all available connections. Requests start queuing, timeouts cascade, and the entire service goes down — often with misleading error messages that point away from the actual cause.

This scenario should demonstrate connection exhaustion with an undersized pool and no statement timeout, then show how proper HikariCP sizing + statement_timeout prevents cascading failures.

Strategies to compare:

  • Undersized pool with no timeout (baseline — reproduces exhaustion)
  • Right-sized pool based on the connection pool sizing formula
  • Pool + statement timeout as a safety net
  • Pool metrics/monitoring to catch problems early

k6 load test should ramp connections beyond pool capacity. Proof: baseline causes connection wait timeouts; fixed config sustains throughput. Follows PRD acceptance criteria AC-1 through AC-5.

Metadata

Metadata

Assignees

No one assigned

    Labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions