Development Roadmap

---

Version Overview

v1.0.0 (current) --> v1.1.0 --> v1.2.0 --> ...
   │                    │          │
   ▼                    ▼          ▼
 Stable              Data       Scale &
 Release           Governance  Performance

---

v1.0.0 -- What's Included

go-mink v1.0.0 is the first stable release, providing a complete toolkit for building event-sourced applications in Go.

Core Event Store

• Event store with append-only storage and optimistic concurrency
• PostgreSQL adapter (production) and in-memory adapter (testing)
• JSON serialization with event registry
• Stream management with version constants (AnyVersion, NoStream, StreamExists)
• Filtered feed reads: LoadEventsFromPositionFiltered — read the global feed by indexed axis (event type / stream id / category), for audit & introspection tooling

Aggregates

• Aggregate interface and AggregateBase implementation
• Event application pattern with SaveAggregate / LoadAggregate

Command Bus & CQRS

• Command interface with validation support
• CommandBus with middleware pipeline
• Type-safe generic handlers (NewGenericHandler, NewAggregateHandler)
• Built-in middleware: Validation, Recovery, Logging, Metrics, Timeout, Retry, Correlation ID, Causation ID, Tenant, Idempotency
• Idempotency store (PostgreSQL and in-memory)

Projection Engine & Read Models

• Inline, async, and live projections
• ProjectionEngine with worker pool, checkpointing, and rebuilding
• Async worker resilience: transient-vs-poison error classification (AsyncOptions.ErrorClassifier / DefaultErrorClassifier / ErrTransient) so infrastructure blips retry independently of the poison budget, plus Faulted-worker supervision (AsyncOptions.RestartPolicy, ProjectionEngine.Restart) with checkpoint-safe resume and a push-based WithProjectionStateObserver fault alert
• ReadModelRepository[T] generic interface with fluent query builder
• Event subscriptions: SubscribeAll, SubscribeStream, SubscribeCategory
• Catch-up and polling subscriptions with event filters

Saga / Process Manager

• Saga interface and SagaBase implementation
• SagaManager for orchestration with event correlation and idempotency
• Compensation handling for rollback workflows
• Operator-initiated safe re-drive of a settled saga (RetrySaga, plus ResumeStalled / RetrySagasByType) — concurrency-safe, idempotent, and auditable via WithSagaRetryObserver
• SagaStore with PostgreSQL and in-memory implementations

Outbox Pattern

• OutboxStore interface with PostgreSQL and in-memory implementations
• EventStoreWithOutbox wrapper with automatic routing
• OutboxProcessor background worker with polling, retry, and dead-letter handling
• Built-in publishers: Webhook, Kafka, SNS
• Atomic event+outbox writes (PostgreSQL)
• Prometheus metrics integration

Event Versioning & Upcasting

• Upcaster interface for raw byte-level transformations
• UpcasterChain with thread-safe registry and gap validation
• UpcastingSerializer decorator
• SchemaRegistry with compatibility checking
• Automatic upcasting on load, schema stamping on save
• Zero overhead when not configured

Field-Level Encryption

• Envelope encryption with AES-256-GCM
• Providers: Local (testing), AWS KMS (production), HashiCorp Vault Transit (production)
• Per-tenant encryption keys via WithTenantKeyResolver()
• Encryption metadata in Metadata.Custom (no DB schema changes)
• Zero overhead when not configured

GDPR & Data Governance

• Crypto-shredding via key revocation (encryption.Revocable / RecoverableRevocable, WithDecryptionErrorHandler())
• DataExporter — right to access / portability (Article 15 / 20)
• DataEraser — right to erasure (Article 17): subject discovery, sibling-store erasure (audit/saga/snapshot/outbox/idempotency), blast-radius guard, and a verification certificate
• SubjectResolver + subject index / BackfillSubjectIndex
• RetentionManager — configurable retention (shred / redact / anonymize); optionally checkpoint-resumable + per-run bounded so a scheduled sweep's cost tracks the retention window, not total history (WithRetentionCheckpoint / WithRetentionMaxScan)
• Anonymizer — deterministic one-way pseudonymization
• mink gdpr CLI — discover / verify / erase-plan / retain
• See the GDPR & Data Governance guide

Observability

• Prometheus metrics middleware (middleware/metrics)
• OpenTelemetry tracing middleware (middleware/tracing)
• 17+ Prometheus metrics for commands, event store, projections, outbox

Serializers

• JSON serializer (built-in)
• MessagePack serializer (serializer/msgpack)
• Protocol Buffers serializer (serializer/protobuf)

CLI Tool

• mink init -- Project scaffolding
• mink generate -- Code generation (aggregate, event, projection, command)
• mink migrate -- Schema management (up, down, create, status)
• mink projection -- Projection management (list, status, pause, resume, rebuild)
• mink stream -- Stream inspection (list, events, stats, export)
• mink events -- Global-feed inspection by indexed axis (type / stream / category), table or --json
• mink diagnose -- Health checks and diagnostics
• mink schema -- Event schema generation
• mink gdpr -- Data governance (discover, verify, erase-plan, retain dry-run)

Testing Utilities

• BDD-style fixtures (testing/bdd) -- Given/When/Then
• Event assertions and diffing (testing/assertions)
• Projection test helpers (testing/projections)
• Saga test fixtures (testing/sagas)
• PostgreSQL test containers (testing/containers)

---

Future Roadmap

v1.1.0 -- Data Governance

• Data export (GDPR right to access)
• Data erasure (GDPR right to be forgotten) — DataEraser + sibling-store erasure
• Audit logging middleware
• Data retention policies with configurable rules
• Time-travel queries (load aggregate at timestamp or version)

v1.2.0 -- Scale & Performance

• Multi-tenancy strategies (shared table, schema-per-tenant, database-per-tenant)
• Snapshot store interface and adapters (PostgreSQL, Redis)
• Additional adapters: MongoDB event store, Redis read models
• Performance optimization: connection pooling, batch loading, benchmarking suite

Future

Reliability

• Dead letter queue improvements (partial: async projections support OnPoisonEvent skip/dead-letter and transient-error classification)
• Circuit breaker pattern
• [~] Graceful degradation (async projection fault supervision landed: RestartPolicy with checkpoint-safe restart, manual Restart, and WithProjectionStateObserver fault alerting)

Advanced Projections

• Projection versioning
• Projection dependencies
• Projection analytics

Production Tooling

• Configuration validation
• Migration rollback improvements
• Backup/restore utilities

Community Adapters

• DynamoDB, ScyllaDB, CockroachDB
• ClickHouse (analytics projections)
• Elasticsearch (search projections)
• NATS JetStream

Ecosystem

• gRPC API for remote event store
• GraphQL read model API
• Admin UI dashboard
• Kubernetes operator
• Event sourcing debugger (visual timeline)

---

Contributing

How to Contribute

1. Pick an issue - Check GitHub issues labeled good first issue
2. Discuss - Comment on the issue or join Discord
3. Fork & Branch - Create feature branch from develop
4. Implement - Follow code style guidelines
5. Test - Add unit and integration tests
6. PR - Submit pull request targeting develop

Priority Areas for Contributors

Area	Difficulty	Impact
Documentation	Easy	High
Test coverage	Easy	High
MongoDB adapter	Medium	High
Redis adapter	Medium	Medium
Performance benchmarks	Medium	Medium
Data export	Medium	Done
Audit logging	Medium	Done
DynamoDB adapter	Hard	Medium
Snapshot support	Hard	High

Code Style

• Follow standard Go conventions
• Use gofmt and golangci-lint
• Write table-driven tests
• Document public APIs
• Keep functions focused and small

---

Success Metrics

Adoption Targets

Milestone	Target
v1.0.0 release	500 GitHub stars
v1.0.0 + 6 months	10 production users
v1.1.0 release	1,000 GitHub stars
v1.2.0 release	25 production users

Quality Metrics

• Test coverage > 90%
• Zero critical bugs in release
• Documentation for all public APIs
• Response to issues < 48 hours

---

Get Involved

• GitHub: go-mink.dev
• Discord: discord.gg/go-mink-go
• Twitter: @go-mink_go
• Blog: blog.go-mink-go.dev

---

Let's make Event Sourcing accessible to every Go developer!