kernablekernable

For event-driven systems · Outbox · Broker · CQRS

Delete your outbox, broker and sync jobs.

If keeping events flowing takes more time than building features: kernable stores, orders and delivers every change in one process, and every read sees it.

[ What it replaces ]

Six pieces to keep in step. Or one.

Event-driven systems are usually assembled from several pieces, each built, run and kept in step by your team.

What you would otherwise build and run

  • Databasecurrent state
  • Outboxreliable publishing
  • Message brokerdelivery
  • Audit tableshistory
  • Sync jobsviews and search
  • Erasure scriptsGDPR requests
6 → 1
kernable – the kernel One process

Stores, orders, signs and delivers every change, keeps the history and erases personal data.

  1. 01Dual writes and an outbox table→
    One atomic commit

    Events are committed atomically into one log; delivery reads from there. Nothing to keep in sync.

  2. 02A broker to run, tune and monitor→
    Delivery built into the kernel

    Every connected service gets the events it may see, straight from the kernel.

  3. 03Consumers that see events out of order→
    A fixed order for every listener

    Read models first, then reactors, each in the order the events were committed.

  4. 04Polling until the read side catches up→
    Visible everywhere right after saving

    A submit returns once every listener has processed it, so the next query sees it.

  5. 05A separate pipeline to replay history→
    Rebuild any view from the log

    A read model starts again from the first event, whenever you change it.

[ How it works ]

One command, from submit to every read.

What happens inside the kernel when a service submits a command, step by step.

What happens when submit runs

[ Where kernable is young ]

Know the limits before you plan.

kernable is in preview. For teams coming from a broker, these matter most:

  • One primary per store. A standby with a witness takes over within seconds; there is no clustering beyond that pair yet.
  • No compaction. The log keeps every event; size it for your volume.
  • The API may change before 1.0. Early access comes with the changes announced and explained.
[ Consulting ]

Move one domain at a time.

We help you pick the first domain, run it on kernable next to your existing system, and plan the rest from what you learn, without a big-bang migration.

Artur Susdorf
Artur Susdorf Author of kernable

Twenty years of shipping software, the last seven designing the architecture other engineers build on. Author of comby, the event sourcing and CQRS framework that carries multiple commercial platforms in regulated markets. kernable brings the same model to teams that work in any language.

LinkedIn →

[ Early access ]

Build on kernable before 1.0.

kernable is in private preview. Tell us what you're building, and we'll send you the binaries and the docs.