Data Snapshots & Rollback

Lightweight snapshots with instant rollback. Reset your database to a known state in seconds.

---

The Problem

After test runs or experiments, resetting the database to a known state is painful and slow.

The Solution

Point-in-time snapshots that restore in seconds, not hours.

Timeline View:

──●────────●────────●────────●────────●──▶ time
  │        │        │        │        │
Sync    Sprint   Sprint   Sprint   Current
#1      End #1   End #2   End #3

Any snapshot can be restored in seconds.

---

Features

• Auto-snapshot after each sync job
• Manual snapshot creation (one-click)
• Scheduled snapshots (every sprint end)
• Named snapshots ("before-migration", "clean-state")
• Instant rollback to any snapshot
• Snapshot diff (compare what changed)
• Retention policies (keep last N, expire after X days)

---

Use Cases

Post-Test Reset

Run E2E tests → DB is dirty → Rollback → Clean state

Sprint-Based Refresh

Every 2 weeks: Auto-snapshot → Fresh sync from prod Problem during sprint? → Rollback to sprint start

Debugging Historical Issues

"This bug wasn't there last week" → Restore last week's snapshot → Reproduce → Compare

Safe Experimentation

Create snapshot → Try risky changes → Rollback if bad

---

CLI & API

# CLI Commands
$ phony snapshot create --name "pre-migration"
$ phony snapshot list
$ phony snapshot restore snap_abc123
$ phony snapshot diff snap_abc123 snap_def456

# API
POST /api/v1/projects/{id}/snapshots
GET  /api/v1/projects/{id}/snapshots
POST /api/v1/snapshots/{id}/restore
GET  /api/v1/snapshots/{id}/diff/{other_id}

---

Scheduling (Team+)

• Every sync: Auto-snapshot before & after
• Cron: 0 0 * * FRI (every Friday midnight)
• Sprint-based: Integrate with Jira/Linear
• Retention: "Keep last 10" or "Expire after 30 days"

---

How It Runs

Snapshots are captured and restored by the phony agent in your infrastructure, and stored in your storage — local disk or your own S3/GCS bucket, encrypted with keys you control. Phony Cloud keeps only the catalog: snapshot ids, schema hashes, sizes, and lineage, which powers scheduling, retention policies, and point-in-time restore. Snapshot contents never leave your network. Details: Snapshots Architecture.

---

Storage

• Your storage: snapshots live on local disk or your own S3/GCS bucket; the cloud tracks metadata only
• Incremental: Only store diffs, not full copies
• Compressed: Typical 10-20% of full DB size

---

Tier Limits

Environments spawned from snapshots (CI runs, agent sandboxes, ephemeral databases) are unlimited at every tier — only tracked snapshots are tiered.

Feature	FREE	STARTER	TEAM	BUSINESS
Tracked Snapshots	3	10	50	Unlimited
Incremental Snapshots	✗	✓	✓	✓
Auto-snapshot	✗	✓	Before & after	✓
Scheduled	✗	✗	✓	✓
Snapshot Diff	✗	✗	✓	✓
Point-in-Time Restore	✗	✗	✗	✓
Retention Policy Rules	✗	✗	✗	✓