Skip to content

--load: ramp-up and stepped load to find a server's knee #2841

Closed

Description

@CloneOfAlex

Which client?

All / shared core

The problem

A fixed concurrency or rate shows how a server behaves at one load level. Finding where latency or errors start climbing (the knee) means many separate runs and stitching their summaries together by hand.

Solution you have in mind (optional)

--load-ramp <from>:<to>:<duration> (concurrency or rate) and/or --load-steps 10,25,50,100 --load-step-duration 30, with one summary row per step, and optionally stop at the first step that breaches a threshold.

Alternatives or workarounds you have tried

A shell loop over --load-concurrency values, parsing each JSON summary with jq.

Already built it locally?

No response

Before you submit

  • I searched existing issues and this is not a duplicate.
  • This is a request for the Inspector itself, not for the MCP specification or an SDK.

Activity

  1. cliffhall commented on Oct 29, 2026

    @cliffhall
    Member

    Triage: Priority Low (total 4)

    • Severity 2 — ramped --load; a shell loop over concurrency values is a workaround
    • Urgency 2 — wanted eventually, nothing blocked on it
    • Bonuses: none (no milestone, assignee, epic, engagement or reporter-set priority)

    Board: #28, Status Incoming (awaiting maintainer review — no milestone yet).

  2. added this to the v2.15.0 milestone on Oct 29, 2026
  3. self-assigned this
    on Nov 5, 2026
  4. cliffhall commented on Nov 5, 2026

    @cliffhall
    Member

    Shipped in #2909 (merge 219e605).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

enhancementNew feature requestv2Issues and PRs for v2

Type

No type

Projects

No projects

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions