Roadmap

SimpleDeploy follows a small, focused roadmap. The day-to-day backlog lives in GitHub Issues and the GitHub Project board. The themes below describe direction, not commitments.

Shipped

• Shared Docker network for apps: endpoint services auto-join simpledeploy-public; upstreams resolved by container IP. Ships in v<next>.

Near term (next few releases)

• More backup strategies: ClickHouse, Elasticsearch, generic exec hooks for “run this command, capture stdout”.
• Backup target plugins: GCS, Azure Blob, SFTP.
• SSO: OIDC for the dashboard so teams can drop password rotation.
• Audit log shipping: stream the audit log to syslog or an HTTP sink for compliance.
• Better deploy diff: show what’s about to change before applying.
• CLI quality of life: tab completion improvements, structured --output json everywhere.

Medium term

• Multi-host federation: a thin coordinator so two or more SimpleDeploy instances appear as one in the dashboard. Not full orchestration; just unified visibility and per-host targeted deploys.
• Built-in load balancer driver: Cloudflare and a generic “tell external LB to drain me” hook.
• First-class secrets: a per-app encrypted secret store that exposes values into the compose env without writing to disk in plaintext.
• Plugin loader: a stable interface for compiled-in extensions (custom Caddy modules, custom strategies/targets).
• Docker install feature parity: document one-shot container workflows (init, rotate-secret, db-backup via docker compose run --rm) as first-class, and add an E2E suite variant (make e2e-docker) so the container path stays green on every release.

Long term

• One-click upgrade: dashboard-driven self-upgrade with automatic backup and rollback. Mainly a concern on the apt/tarball paths; on the Docker install, docker compose pull && up -d already covers this.
• App templates marketplace: curated compose files for common stacks.

What we will not do

• Replace Kubernetes. SimpleDeploy is for one to a few hosts.
• Add a hosted control plane. Single-binary, self-host first.
• Lock-in to a specific cloud or registry.

Have a different priority? Vote on issues or open one. Maintainers steer by community signal.