Sync queue and health

Read queue behaviour and respond to health signals quickly.

Last updated:

Use Control Centre to review current per-channel health signals, including recent failures, retries, and last successful sync times.

Queue fundamentals

Each sync action enters a queue with retry metadata and event history.

  • Manual refreshes and scheduled syncs are queued so ChannelWeave can retry safely when a channel is slow or temporarily unavailable.
  • Repeated refresh clicks are coalesced so the same channel is not flooded with duplicate requests.
  • Listing publish attempts keep enough history for interrupted remote publishes to be reviewed and reconciled. Interrupted attempts appear as Needs attention in Listing drafts and block another publish retry until ChannelWeave or an operator reconciles them.
  • Stock-consuming orders reduce available listing quantities. Website listings update immediately; external marketplaces continue through the channel sync queue.
  • Account-deletion and disconnect events are verified before account data is removed or scrubbed.

Health indicators

  • Queue depth
  • Retry count
  • Oldest pending age
  • Failure ratio by channel
  • Last successful sync time
  • Stuck or repeatedly failing jobs
  • Last error summary

Where to look

  • Control Centre: current connection state, last successful sync signals, and recent issue counts.
  • Events & diagnostics: event-by-event detail for channel actions, webhooks, and sync jobs.
  • Listings and Listing drafts: item-level publish and listing state. Listing drafts also shows the latest publish error or attention-required reconciliation message inline, so failed rows do not require opening one by one.
  • Orders: imported order state and downstream fulfilment impact.

Common causes of sync pressure

  • Expired or revoked channel authorisation.
  • Marketplace throttling or temporary provider outage.
  • A publish request that needs user correction before retry.
  • A listing or order payload that is missing required channel data.
  • Repeated manual retries before the previous job has finished.

Incident response

  1. Pause noisy automations.
  2. Fix root cause.
  3. Replay failed jobs.
  4. Verify output parity.