Skip to content

E10: Observability and SRE for Sith itself #28

Description

@gnanirahulnutakki

Goal. make the hub — the crown jewel — observable and operable: metrics, tracing, and structured logs about Sith's own behavior, SLOs with error budgets, and the hardening a fleet-wide control plane demands.

Phase / depends. the hardening posture is day-one; the surfaces mature P1 → P3. Depends on: E9 (deployment) and touches every other epic. This epic observes Sith itself — it does not store other systems' telemetry (that would be a telemetry lake, out of scope). The hub is the highest-value target and largest blast radius in the estate, so its own operability and hardening are first-class.

Features

  • F10.1 — metrics
  • F10.2 — distributed tracing
  • F10.3 — structured logging
  • F10.4 — SLOs and alerting
  • F10.5 — crown-jewel hardening

Exit criteria

  • Sith exposes metrics, traces, and structured sanitized logs about its own behavior, correlated by intent id, with no secret leakage and no other-system series retention.
  • SLOs with error budgets cover read freshness, dispatch success, and PDP latency; alerts cover reliability and security conditions.
  • Crown-jewel hardening is in place: KMS/HSM key custody, RLS + least-privilege hub identity, signed images + SBOM, pinned/verified addons.

Source: docs/EPICS.md, docs/SITH-NOTION.md · part of the Sith implementation backlog (see the master roadmap issue).

Metadata

Metadata

Assignees

No one assigned

    Labels

    epicEpic tracking issuephase-1Phase 1: read-only federation

    Type

    No type

    Fields

    No fields configured for issues without a type.

    Projects

    No projects

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions