Miren-hosted runners

[phinze]

Once #1 ships, you will be able to run workloads on many different nodes, but you still have to figure out where and how those nodes run.

With Miren-managed runners, you will be able to opt-in to having runners execute inside Miren Cloud. This has a number of tangible benefits, one of which is an easy onramp to elastic capacity without tethering your entire cluster to a managed compute offering.

You can run a control plane where it makes sense for you, say in a data center that you own and manage, and then use Miren-managed runners to run the the production version of an app while utilizing your local compute capacity for test and staging.

The name of the game here is flexibility. We aim to give you a set of solutions with tradeoffs that fit your needs.

Before (self-managed runners only):

Assumes multi-node runners is complete.

graph TB
    subgraph "Your Infrastructure"
        C[Coordinator]
        R1[Runner]
        R2[Runner]
    end

    C --> R1
    C --> R2

Loading

After (Miren-hosted runners):

graph TB
    subgraph "Your Infrastructure"
        C[Coordinator]
        R1[Runner - test/staging]
    end

    subgraph "Miren Cloud"
        R2[Runner - production]
        R3[Runner - production]
    end

    C --> R1
    C --> R2
    C --> R3

Loading