Access by request

Put one workload on a budget

Name the workload you want to test first. We run it under your policy and send back the ledger.

  1. 01

    A platform engineer replies

    Within two working days, by email.

  2. 02

    You write the policy together

    A quality floor and two ceilings, in one 30-minute call.

  3. 03

    We run it and send the ledger

    What each route delivered and consumed.

Rather write?access@joulis.io

Tell us about the workload

About 2 minutes

Use a work email, so the reply reaches your team.

Add your name.

Add the company you work for.

Monthly requests

One workload is enough. We ask for sample requests on the call.

Next: a platform engineer replies within two working days.

What comes back from one run

Your policy and your requests go in. A ledger of every route comes back, with any exception named.

  1. 01Your policy

    A floor, two ceilings and a fallback, written with you on the call.

    policy
    support-summaries
    quality
    ≥ 89/100
    latency
    ≤ 1.8 s p95
    energy
    ≤ 60 Wh / 200 est.
    fallback
    gannet-l
    no route fits
    report, do not run

  2. 02Your workload

    Requests from your own traffic, with the same input and output limits on every route.

    • Requests200your traffic
    • Limitssameall routes
    • Routed0not yet

  3. 03The ledger, back to you

    Every route, its score, its latency and its energy estimate. Shown: Dispatch-200.

    • Rubric90/100fixed rubric
    • p951.4 smeasured
    • Energy50 Whest.