Skip to content

Roadmap

import { Callout } from ‘@astrojs/starlight/components’;

FourEyes would rather under-promise. These are real gaps, not “enterprise” features dressed up as differentiators. If something matters to you, it probably matters to the project — open an issue.

Today the platform runs on SQLite (relational + time-series, pure Go) or DuckDB (CGO opt-in). The PostgreSQL + ClickHouse “scale” backend is the next storage tier and is not shipped. FOUREYES_STORAGE_MODE=scale is reserved.

The alert engine currently has a single in-process sink (LogSink). Routing to email, Slack, PagerDuty, Alertmanager, and webhooks is a planned increment that adds new sinks. The incident/resolve model and the /v1/events/alerts stream are already in place.

A default-org scope is the current stand-in for full multi-tenant RBAC. Note that /metrics is cross-org by design and must be gated before any multi-tenant deployment.

A one-line installer ships for macOS and Linux (install.sh) and Windows (install.ps1). It detects your OS/arch, verifies the SHA256 checksum, and installs to /usr/local/bin (or $ProgramFiles\FourEyes). Missing: native apt/dnf packages and a documented air-gapped bundle (mirroring the binaries offline works today but isn’t packaged).

The voice probe is signaling-only today (SIP OPTIONS/REGISTER with digest auth). Full RTP / E-model MOS validation needs a far-end peer and isn’t shipped.

BGP ships receive-only session monitoring. The per-prefix route-event log, RIB dump, and AS-path topology UI are planned follow-ups.

Event correlation groups simultaneous incidents by shared target / hop / ASN. The full “mine vs theirs” verdict (which needs an AS-ownership / “my prefixes” model) is forward-compatible but not yet populated.

The atomic swap + re-exec self-update path is linux/macOS. On Windows it’s a measured no-op in this increment.

Pointers for understanding the system: [Architecture](/concepts/architecture/) for the mental model, [Probe types](/concepts/probes/) for what the registry looks like, and [Security model](/concepts/security/) for the trust boundaries.