Skip to content

Draft: SLA and incident-response commitment for hosted AMS #5227

Description

@JSONbored

Problem

ORB's #4936 tracks a published SLA and incident-response commitment for ORB's hosted fleet. AMS has no generic equivalent - the only incident-response work currently tracked for AMS (Rent-a-Loop's #5137) is specific to the rented-loop product and its own harm model, not a general operational commitment for a future hosted AMS.

Area

AMS / Incident & SLA readiness

Proposal

Draft an SLA and incident-response commitment appropriate to a future hosted AMS: an uptime target, a response-time commitment, and an escalation path. This is a business/operational commitment gittensory would be making to hosted-AMS customers, separate from - and composing with, not replacing - Rent-a-Loop's own harmful-loop incident process.

Deliverables

Resources

Boundaries

  • This is a business/legal commitment, not a technical design - but it still requires maintainer sign-off before it is treated as adopted, since it commits gittensory to an external obligation toward hosted-AMS customers.
  • Must not be published, referenced in marketing/docs, or represented externally as a live commitment until explicitly approved by the maintainer.
  • Must not duplicate or override Rent-a-Loop's Post-merge incident process for an already-landed rented-loop PR found harmful #5137 incident process - the two should be shown to compose, with Post-merge incident process for an already-landed rented-loop PR found harmful #5137 as the product-specific instance of the general commitment drafted here.
  • No numeric SLA target (uptime %, response-time bound) in this draft may be treated as final without maintainer approval, since these numbers create binding obligations.

Metadata

Metadata

Assignees

Labels

maintainer-onlyOwner-only work — yields no Gittensor points.roadmapOn the Wave-2 agent-layer roadmap board (project 9)

Projects

No projects

Relationships

None yet

Development

No branches or pull requests

Issue actions