docs(design): Debezium-compete CDC parity roadmap#2643
Merged
Conversation
Sequences a Debezium-compete epic (DBZ-0..10) into the existing roadmap phases. Grounded in a verified per-connector CDC-mechanism audit (2026-07-22): log-based parity on Postgres/MySQL/Mongo/Vitess, trigger-based on Oracle/SQL Server/Db2, Spanner/Cassandra CDC missing. Two tracks — KC-wrapper coverage (fast credibility) and native log-based depth (the moat) — gated by a CDC correctness suite. Includes a proposed Phase-2 CDC section re-scope. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
… reality, sequencing, UX) Four independent reviews (2 technical, 2 sequencing/UX) found the doc treated the Kafka Connect wrapper as greenfield when conduit-kafka-connect-wrapper already exists and passes DebeziumPgSourceIT in CI. This revision: - Rewrites Track A / DBZ-1 / DBZ-7 around validating and extending the existing wrapper instead of building one; retracts the earlier "async OffsetBackingStore trap" claim (offset/ack ordering is actually correct); names the strategic inversion that the wrapper only proves Postgres (least-needed) while MySQL/Oracle/SQL Server/Db2 need an unwired schema-history store. - Flips the DBZ-3/DBZ-4/DBZ-5 dependency inversion so the flagship (Postgres) and second implementation (MySQL/Mongo) land before shared SDK plumbing is extracted, and makes the critical-path sentence agree with the dependency table. - Tightens DBZ-2 to SIGKILL-explicit crash tests covering the wrapper's ack/offset path plus an invariant-6 DDL-mid-stream check. - Adds an Operator UX and observability section (API surface, CLI parity, DLQ Tier-1 gap inheritance, runbook stub) and smaller fixes (JVM failure-mode reframe, embeddability positioning note, tx-metadata honesty, Track-A Phase-3 checkpoint, SQL Server README mislabel). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Restate the embeddability differentiator as a settled vs-Debezium comparison claim rather than an open question: it names Debezium Server and Debezium Engine directly, and grounds "already shipped" accurately (conduit run ships the standalone binary today; libconduit is the embedding path, Go embedding working today with the stable API as its embed-v1 productization). DBZ-1: the doc makes no "a few instructions" crash-window claim to correct, so DBZ-1's detailed kill -9 crash-safety test is left to its own later PR. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_015GQFzakPShAYj8CcwajYDD
devarismeroxa
added a commit
to ConduitIO/conduit-site
that referenced
this pull request
Jul 24, 2026
Publishes the CDC coverage/mechanism/maturity audit from the internal Debezium-compete roadmap (ConduitIO/conduit#2643) as a public, honest comparison page: leads with embeddability (conduit run + the new libconduit B1 Go embedding API subsume Debezium Server/Engine's standalone-process use case, with the scope caveat locked in prose), an accessible <table> matrix (caption + th scope, icon+text status, never color alone), a "Choose Debezium if..." section at equal visual weight to the parity claims, and a KC-wrapper honesty split (Debezium-Postgres is CI-tested via DebeziumPgSourceIT; MySQL/Oracle/SQL Server/Db2 via the wrapper are unwired, scoped future work). Crash-safety is explicitly not claimed yet — links to issue #2672 and the roadmap doc as a verification-in-progress placeholder, pending DBZ-1. Every factual claim links to connector source/README or a named CI job, not to the internal roadmap doc. Claude-Session: https://claude.ai/code/session_015GQFzakPShAYj8CcwajYDD Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Tier 3 (design doc). Not for immediate merge — opened for DeVaris review/edit.
What this is
A sequenced Debezium-compete epic (
DBZ-0..10), placed into the existing roadmap phases rather than standing apart. Grounded in a verified per-connector CDC-mechanism audit (2026-07-22) — reading each connector's actual mechanism, not its description.The honest picture it's built on
conduitio-labs).The plan
Two parallel tracks, gated by a CDC correctness suite:
Critical path to credible messaging: DBZ-0 (honest
vs Debeziumpage) → DBZ-1 (KC-wrapper PoC) → DBZ-2 (correctness suite) → DBZ-3 (Postgres to real parity).Placement
Full sequencing table maps each workstream to a Phase 1/2/3 slot with dependencies. Includes a proposed Phase-2 CDC section re-scope for
ROADMAP.md(to apply once the roadmap-reconciliation PR #2638 settles, to avoid a conflicting edit).Review asks: is the Track-A-vs-B sequencing right for the near-term credibility goal, and is the proposed ROADMAP delta the shape you want before I touch
ROADMAP.md?🤖 Generated with Claude Code