FleetAMP roadmap
FleetAMP is building a vendor-neutral telemetry control plane in deliberate layers. Milestones describe direction rather than promised dates or commercial support guarantees.
| Milestone | Status | Scope |
|---|---|---|
| v0.1.0 — Foundation | Released | OpAMP inventory, lifecycle, stable identity, persistent groups and deployments, rollback, first-admin setup, TLS and Linux packaging |
| v0.2.0 — Governed fleet management | Release candidate | Group ownership, scoped RBAC, governed configuration sections, validation, approval and expiry, deployment safeguards, automatic rollback, drift policy and audit |
| v0.3 — Guided instrumentation | Planned | Technology/platform guides, reusable Telemetry Blueprints, configuration modules and destination profiles |
| v0.4 — Telemetry governance | Planned | Telemetry contracts, semantic-convention governance, policy as code and telemetry-quality controls |
| v0.5 — Intelligence | Planned | Cardinality, volume, cost and PII signals with explainable recommendations |
| v0.6 — Discovery | Planned | Workload and service discovery, eBPF/OBI integrations and instrumentation recommendations |
| Later — Platform scale and agentic operations | Exploration | PostgreSQL and HA, canaries, package upgrades, Kubernetes packaging, more adapters, topology, profiles, AI-assisted operations and an audited MCP interface |
| v1.0 goal | Future | Stable compatibility, recovery, upgrade, security and support expectations backed by production-scale validation |
v0.2.0 release boundary
The v0.2.0 release is complete when the governed loop is dependable:
- A Collector is identified and placed in a controlled group.
- An authorized owner edits permitted configuration sections.
- FleetAMP reconstructs and validates the complete configuration.
- A separate Admin reviews an immutable request and change comparison.
- FleetAMP deploys to the snapshotted targets and records status.
- Failed delivery triggers one rollback attempt to the previous working version.
- Drift and administrative activity remain visible and auditable.
Partial conf.d delivery, canary stages, Collector binary upgrades, OIDC,
PostgreSQL/HA and production-scale certification are intentionally outside this
release boundary.
Design principles
- Keep FleetAMP independent from telemetry storage vendors.
- Prefer upstream Collectors and explicit adapters over a custom distribution.
- Separate desired state, approval state and observed runtime state.
- Make ownership, policy decisions and deployment outcomes auditable.
- Show future direction without presenting unfinished features as available.
Proposals and implementation progress are tracked in the FleetAMP repository.