fix: harden tracker and uptime delivery

[izadoesdev]

Summary

• bound tracker pixel retries, image loads, and cancellation; pixel unloads retain the normal fallback path
• make uptime delivery durable: checkpoint completed probes on the source job, relay them through BullMQ, and require acknowledged Kafka sends
• keep the relay event ID inside the worker and delivery queue; strip it before Kafka/Vector so the existing ClickHouse schema remains unchanged
• refresh existing schedulers with the durable job options and harden uptime worker shutdown

Scope

Clean replacement for #581 on current staging. It intentionally excludes that branch's stale Basket and Vector changes.

Verification

• bun run lint
• bun run check-types
• focused tracker, uptime, Redis, AI query, database schema, and RPC tests

[cursor]

Bugbot is not enabled for your account, so this pull request was not reviewed.

Enable Bugbot in the Cursor dashboard to get automatic reviews on future PRs.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
dashboard	Ready	Preview	Aug 2, 2026 2:10pm
databuddy-status	Ready	Preview	Aug 2, 2026 2:10pm
documentation	Ready	Preview	Aug 2, 2026 2:10pm

[unkey-deploy]

The latest updates on your projects. Learn more about Unkey Deploy

Name	Status	Preview	Inspect	Updated (UTC)
links (preview)	Ready	Visit Preview	Inspect	Aug 2, 2026 2:09pm

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: 274a03a7-8e10-4301-9874-71913b0aa65c

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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

[dosubot]

📄 Knowledge review

Dosu skipped reviewing this PR because your organization has used its 200 included credits for the month. Your usage will reset on 2026-09-01. To have Dosu review this PR before then, ask your organization admin to upgrade to a pro account.

---

[greptile-apps]

Greptile Summary

This PR makes uptime event delivery durable through a checkpointed BullMQ relay and acknowledged Kafka sends, refreshes scheduler retry options, adds replay-aware reporting queries, and bounds tracker pixel retries and cancellation.

• Adds per-probe event IDs, persisted source-job checkpoints, and a dedicated uptime delivery worker.
• Reconciles existing schedulers with durable retry settings and performs ordered worker shutdown.
• Deduplicates replayed ClickHouse rows in uptime and status-page reporting.
• Bounds pixel image loads and retry delays while retaining the normal unload fallback.

Confidence Score: 4/5

The reporting deduplication should be corrected before merging because it can discard distinct uptime checks that happen to have identical stored values.

The new event ID provides a unique probe identity only until ingestion, where it is dropped; downstream queries consequently deduplicate by mutable observation values and cannot distinguish an exact replay from two genuinely separate identical probes.

Files Needing Attention: packages/ai/src/query/builders/uptime.ts and packages/rpc/src/routers/status-page.ts

Important Files Changed

Filename	Overview
apps/uptime/src/worker.ts	Adds checkpoint replay, a dedicated delivery worker, payload validation, and ordered durable handoff.
packages/redis/uptime-queue.ts	Defines the delivery queue, deterministic event job IDs, and long-lived fixed retry options.
apps/uptime/src/lib/producer.ts	Reworks Kafka producer lifecycle to share connections, require acknowledgements, and surface send failures.
packages/ai/src/query/builders/uptime.ts	Adds full-row replay deduplication that can also collapse distinct probes whose stored values coincide.
packages/rpc/src/routers/status-page.ts	Applies the same value-based replay deduplication to public status calculations.
packages/tracker/src/plugins/pixel.ts	Bounds image loads and retries, integrates cancellation, and preserves fallback behavior on unload.
apps/uptime/src/index.ts	Starts both workers and drains them in source-before-delivery order with fatal exit-code preservation.

Sequence Diagram

sequenceDiagram
  participant S as BullMQ source job
  participant P as Uptime probe
  participant D as BullMQ delivery queue
  participant K as Kafka
  participant V as Vector
  participant C as ClickHouse
  S->>P: Run probe
  P-->>S: UptimeData with event_id
  S->>S: Checkpoint event in job data
  S->>D: Add deterministic delivery job
  D->>K: Send with all-replica acknowledgement
  K->>V: analytics-uptime-checks
  V->>C: Insert known uptime columns
  Note over V,C: event_id is dropped
  C-->>C: Reporting applies value-based LIMIT 1 BY

Loading

_{Reviews (1): Last reviewed commit: "fix(uptime): durably relay completed che..." | Re-trigger Greptile}