refactor(ingest): per-table batching + pipelined flushes

[EricAndrechek]

Summary

Per-table batching

Related Issues

Closes #191

[coderabbitai]

Note

Reviews paused

It looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the reviews.auto_review.auto_pause_after_reviewed_commits setting.

Use the following commands to manage reviews:

• @coderabbitai resume to resume automatic reviews.
• @coderabbitai review to trigger a single review.

Use the checkboxes below for quick actions:

• ▶️ Resume reviews
• 🔍 Trigger review

📝 Walkthrough

Walkthrough

This PR refactors ingestion dispatch to parse messages once, route them by table into lazily-created per-table channels, and run independent per-table batching state machines with configurable size and timeout. It also replaces shared E2E fixtures with generated per-suite ClickHouse tables and updates tests, setup, docs, and changelog.

Changes

Per-Table Batching Pipeline

Layer / File(s)	Summary
Batching Configuration and Data Contracts internal/ingest/worker.go	parsedMsg gains tableName. IngestWorker adds maxBatch, maxWait, and ackWg; package defaults and consumer tuning constants are centralized.
Message Parsing and Dispatch Loop internal/ingest/worker.go	parseMsg centralizes envelope parsing (malformed JSON is logged and double-acked). StartIngestWorker starts dispatchLoop which demultiplexes parsed messages into lazily-created, bounded per-table channels and defines shutdown order.
Per-Table Batching State Machine internal/ingest/worker.go	New tableLoop/tableBatcher accumulates per-table rows, flushes on maxBatch or maxWait, enforces one in-flight bulk flush per table, coalesces triggers arriving during a flush, and performs final synchronous drain/flush on channel close.
Acknowledgment and Shutdown Ordering internal/ingest/worker.go	Background double-acks are scheduled on ackWg; handleSuccess uses ackWg for ack goroutine lifetime. Shutdown waits table loops first, then drains ackWg.
Tests: parse, flushTable, tableBatcher, dispatch regressions internal/ingest/worker_test.go	Test harness updated to wait on ackWg. Adds TestParseMsg, rewires flush tests to flushTable via parseAll, adds tableBatcher unit/state-machine tests, and adds dispatch-loop regression tests for per-table isolation and partial-batch correctness.

E2E: Per-Suite Table Isolation & Setup

Layer / File(s)	Summary
E2E table helpers and DDL tests/e2e/sdk/tables.ts	Adds TableKind/KINDS, TABLE_DDL generators, SUITES, suiteTables, tableName, and allTableSpecs() to compute per-suite physical table names and DDL.
Global E2E setup and bootstrap tests/e2e/sdk/setup.ts	Replaces fixture application with createTables() that POSTs generated DDL for allTableSpecs(), refreshes schema against generated names, and bootstraps dynamic policies for those tables.
E2E test suites tests/e2e/sdk/*	Multiple E2E tests updated to import suiteTables("<suite>") and use T.clicks/events/users instead of hard-coded table names; admin tests snapshot/restore baseline policy; DLQ/timing checks and polling windows adjusted.
Docs & Changelog docs/src/content/docs/ingest-pipeline.md, docs/src/config/sidebar.ts, AGENTS.md, CHANGELOG.md	Adds contributor-facing ingest pipeline deep-dive, registers the “Ingest Pipeline” sidebar entry, documents E2E per-suite isolation rules in AGENTS.md, and updates changelog describing per-table batching and E2E setup changes.

🎯 4 (Complex) | ⏱️ ~45 minutes

---

Note

🎁 Summarized by CodeRabbit Free

Your organization is on the Free plan. CodeRabbit will generate a high-level summary and a walkthrough for each pull request. For a comprehensive line-by-line review, please upgrade your subscription to CodeRabbit Pro by visiting https://app.coderabbit.ai/login.

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[gemini-code-assist]

Summary of Changes

Hello, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request refactors the ingestion worker to move from a global batching strategy to a per-table batching architecture. By isolating batches per table, the system ensures that high-volume traffic is not delayed by the batching or timing constraints of low-volume tables. This change also introduces pipelined flushes to maximize throughput by allowing concurrent batch accumulation and ingestion.

Highlights

• Per-table Batching: Refactored the ingestion worker to use per-table batching instead of a global batch, preventing cross-table contamination where low-volume tables could strand events from high-volume tables.
• Pipelined Flushes: Implemented a per-table loop with pipelined flushes, allowing a new batch to accumulate while a previous one is in-flight, improving throughput and reducing latency.
• Improved Reliability: Added a new test case to reproduce and verify the fix for the cross-table contamination issue identified in refactor(ingest): per-table batching + pipelined flushes; unblock per-table-per-test e2e isolation #191.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize the Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counterproductive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request refactors the ingest worker to use per-table batching and pipelined flushing, resolving an issue where low-volume tables could delay high-volume tables. A new test is added to verify this behavior. However, a critical issue was identified in tableLoop where exiting immediately on ctx.Done() bypasses draining buffered messages in the table channel, which can lead to message loss or unnecessary redeliveries on shutdown.\n\nVerdict: Iterate

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Updated (UTC)
❌ Deployment failed View logs	wavehouse-docs	c27550a	Jun 03 2026, 03:17 PM

[taitelee]

Looks good. Per-table batching refactor, wait groups for graceful shutdown, and load distribution using goroutines and channels for ingest. Important to check docs for wait group logic.