Skip to main content
EventDBX is fundamentally built around these ideas: append-only truth, reproducible state, and integrations that never block writes. This page is the high-level compass; jump into the deep dives when you need field-by-field details or CLI examples.

​
Flexibility

  • Start in default mode so teams can append JSON payloads without schema friction, then switch to strict once data shapes stabilise.
  • Use dbx checkout to spin up isolated domains (or delete them) and persist remote settings for replication. This keeps experiments, bounded contexts, and environments walled off.
  • Feed each read model only the slices it needs by tailoring plugin payload selectors (including the extensions-only mode for metadata-only fan-out) and overriding them ad hoc with dbx plugin replay --payload-mode ….
Deep dive → Flexibility

​
Data integrity

  • Per-aggregate Merkle trees and deterministic snapshots keep proofs cheap to recompute and compare across nodes.
  • Schemas use explicit column definitions with built-in types, formats, and validation rules (required, length, range, regex) so intent stays unambiguous.
  • Toggle dbx restrict between permissive and strict modes to match your environment without rewriting history. Auto-snapshots fire at configurable thresholds; create point-in-time snapshots on demand when you need a checkpoint.
Deep dive → Data integrity

​
Security

  • Encrypt payloads, snapshots, and tokens at rest with the DEK; enforce Noise handshakes in transit so credentials never cross plaintext.
  • Append-only storage plus Merkle roots make the log immutable and tamper-evident.
  • Generate Ed25519 tokens per role with fine-grained scopes (TTL, write limits), rotate them regularly, and keep authentication passwordless by design.
Deep dive → Security

​
Auditability

  • Every aggregate keeps an append-only history plus deterministic snapshots for quick reads.
  • Export proofs to auditors or partner teams without replaying production traffic; include Merkle roots with exported snapshots.
  • Stream or plugin-replay events with precise payload modes to retrace every change, then use CLI helpers to archive, restore, and verify aggregate timelines.
Deep dive → Auditability

​
Extensibility

  • Build plugins that emit just the payload slices downstream systems expect, and let the queue absorb backpressure so writes stay fast.
  • Extend schema rules, column types, transports, or plugin payload modes without stopping the daemon.
  • Bridge EventDBX into existing buses, CDC pipelines, serverless workers, or ad hoc replays—plugins operate asynchronously and can be toggled per domain.
Deep dive → Extensibility