Platform Reliability & Background Processing
Strengthen SlabIQ’s background processing, marketplace jobs, retries, error handling and deployment reliability so long-running operations finish predictably and failures are always visible.
- Status
- In Progress
- First discussed
- Development started
- Last updated
Overview
SlabIQ increasingly depends on background workers for imports, catalog processing, marketplace publishing, image processing, sync operations and other long-running tasks. This initiative focuses on making those systems resilient, observable and safe as SlabIQ scales.
Why we're building this
A marketplace action or background task must never silently fail, appear successful when it failed, or leave the user guessing about what happened.
Current work
Improving worker reliability, failure classification, retry behavior, job visibility and deployment consistency.
Completed work
- Publishing results are classified as accepted, rejected or needs checking, so an accepted listing is never retried into a duplicate (2026-09-18)
- Interrupted batch identification runs are detected and can be resumed (2026-09-27)
- Background services reconnect on their own after a brief infrastructure outage (in production by 2026-10-20)
- Incoming Shopify orders are saved on arrival and processed with automatic retries (2026-10-21)
- A CardTrader connection pauses only on a real sign-in rejection and resumes after a passing test (2026-10-21)
- Internal tools trace what happened to each card in a batch (2026-10-21)
- Automated health checks across 13 services, with stuck-card and missed-schedule detection, published on the status page (2026-10-27)
What's next
Better job history, stronger health checks, automatic recovery, marketplace outage handling and clearer admin diagnostics.
Milestones
- Completed
Background worker architecture in place
- Completed
Marketplace/background processing separated into dedicated workers
- In Progress
Centralized error handling
- In Progress
Retry/recovery improvements
- In Progress
Job-state accuracy
- Completed
Better admin diagnostics
- In Progress
Central job history
- In Progress
Automatic stalled-job detection
- In Progress
Marketplace circuit-breaker handling
- Completed
Automated service health monitoring
- In Progress
Consistent releases across the website and background workers
What “done” means
The outcomes this item has to deliver. Each is marked as it is confirmed.
A background task that failed is never shown as successful
Partly met
Service health is checked automatically and published without manual updates
Met
Stuck work is detected without waiting for a user report
Partly met
A release updates the website and all background workers together
Not met yet
Development updates
No development updates have been posted yet.