ROADMAP.md

Meldbase roadmap and production-readiness criteria

This checklist is an evidence index, not a product-support claim. Checked items have implementation, automated-test, or operational-documentation evidence in this repository. Unchecked items need further implementation, target-environment evidence, or an explicit product-scope decision.

Meldbase should establish a workload-specific single-node capacity boundary before considering high availability, horizontal scaling, or managed-service features. See Product scope and production qualification for the supported boundary and the evidence process.

For the detailed query and update operator sequence, see QUERY-AND-UPDATE-ROADMAP.md.

Correctness and resilience

• Long-running concurrent read/write tests.
• Stress coverage for queries, writes, and realtime subscriptions together.
• Crash and abnormal-shutdown recovery tests.
• Equivalent-result coverage for memory and durable backends.
• Index and query correctness after substantial update workloads.
• Index creation, rebuild, failure, and recovery tests.
• Replay, resume, and query-delta coverage without lost or duplicate events.
• Randomized model tests against an independent reference implementation.
• Fault injection for process termination, disk errors, interrupted writes, cancellation, and timeouts.
• Defined fail-stop behavior for unrecoverable database errors.

Backup, restore, and data evolution

• Scheduled backup policy.
• Lower-cost backup option.
• Backup-integrity verification.
• Restore drills in an independent environment.
• Point-in-time recovery capability, or an explicit statement that it is not supported.
• Post-restore consistency for indexes, queries, realtime views, and replay.
• Data export and import tools.
• Storage-format upgrade and older-data migration guidance.
• Defined RPO and RTO expectations.
• Scheduled restore drills rather than backup-file existence checks alone.

Deployment and operations

• Clear single-node deployment documentation.
• Container image or standard binary delivery.
• Consistent configuration-file and environment-variable contract.
• Log output, levels, and host-managed rotation guidance.
• Disk-exhaustion behavior and alerting guidance.
• File permissions, locking, and multi-process misuse handling.
• Health endpoints.
• Readiness and liveness probes.
• Graceful shutdown and in-flight request handling.
• Upgrade procedure.
• Rollback procedure.
• Resource limits and capacity-configuration guidance.
• Troubleshooting guide.
• Production operations runbook.

Performance and capacity boundaries

• Measure read/write throughput across data sizes.
• Measure query P50, P95, and P99 latency.
• Measure concurrent read and write limits.
• Measure realtime subscriber limits.
• Document suitable uses for memory and durable modes separately.
• Measure behavior for large documents, deep nesting, and large arrays.
• Measure index count, index size, and write-maintenance cost.
• Measure behavior before and after query-budget enforcement.
• Measure resource effects across multiple workspaces or collections.
• Publish a capacity guide covering document count, concurrency, subscriptions, memory, disk, and latency.
• Provide repeatable observer and benchmark commands for key workloads.

Security and workspace isolation

• Authentication-key rotation.
• Documented token-expiry and revocation boundary.
• TLS deployment guidance.
• Workspace-isolation tests.
• Query-field authorization tests.
• Filter-operator authorization tests.
• Independent authorization for sorting, aggregation, and result fields.
• Administrator-operation audit events.
• Audit events for authentication and authorization rejection.
• Request rate limiting.
• Request and response size limits.
• Malicious-query and resource-exhaustion stress coverage.
• Sensitive fields excluded from logs, metrics, Explain, and diagnostics.
• External errors do not disclose internal paths, credentials, or business data.

API and protocol lifecycle

• Defined protocol-version policy.
• Wire backward-compatibility tests.
• SDK semantic-versioning policy.
• Stable and experimental markings for Go and TypeScript APIs.
• Stable error-code and error-class documentation.
• Data-format migration policy.
• API deprecation process.
• Release changelog.
• Explicit upgrade compatibility statement.
• Consistent behavior guidance for local, remote, durable, and realtime modes.

Developer experience

• Five-minute introduction.
• Complete CRUD example.
• Realtime-subscription example.
• Query, sorting, and pagination guide.
• Index-design guide.
• Update and delete selection-criteria guide.
• Memory, durable, and remote mode guide.
• Go, Node, TypeScript, and React examples.
• Production-deployment example.
• Backup and restore guide.
• Performance-tuning guidance.
• Common-error troubleshooting.
• Migration boundary for MongoDB and other document databases.

Tooling and integrations

• Stable CLI contract.
• Container image or standard runtime package.
• Administrative API.
• Prometheus metrics.
• OpenTelemetry integration.
• Structured logging integration.
• Data export and import tools.
• Test-database and temporary-database approach.
• CI startup, test, and data-cleanup practices.
• Common authentication-system integration guidance.
• Webhook or event-consumption guidance.

Product boundaries

• Define the primary form: embedded database with an optional realtime server boundary.
• Define memory and durable mode boundaries.
• Define the workspace-isolation boundary without claiming hosted tenancy.
• State the horizontal-scaling boundary.
• State the managed-service boundary.
• State that MongoDB concepts are familiar vocabulary, not wire or driver compatibility.
• Document unsupported capabilities and their alternatives.

Release criteria

Internal use

• Core query and update semantics are stable.
• Durable mode recovers.
• Primary SDKs can integrate.
• Realtime subscriptions are stable.
• Basic backup exists.
• Basic monitoring exists.

Small-scale production consideration

• Complete target-environment concurrency and fault evidence.
• Complete backup and restore drills for the intended deployment.
• Establish performance and capacity boundaries.
• Complete security auditing and workspace-isolation validation.
• Establish upgrade and rollback evidence.
• Establish stable error-code and API-version policy evidence.
• Provide deployment and troubleshooting evidence for the intended environment.

Large-scale production consideration

• High availability and failover.
• Horizontal scaling.
• Node-failure recovery.
• Sharding or partitioning strategy.
• Cross-region replication.
• More complete rate limiting and workspace resource isolation.
• SLA, support, and operating model.

Next evidence priorities

• Crash recovery, backup, and restore drills on the intended storage.
• Concurrency, capacity, and fault injection on target hardware.
• Deployment, upgrade, and rollback evidence for the intended environment.
• Security audit and workspace-isolation validation.
• API, protocol, and error-code lifecycle evidence.
• Developer documentation and real-application examples.
• Re-evaluate high availability, horizontal scaling, and managed service only after the single-node boundary is evidenced.