Troubleshooting
Applications are not reaching Bullhorn
Candidates are applying and consultants are not seeing them. Read this first: the applications are not lost. They are written down on your own site before the candidate is ever told their application succeeded, so what you have is a delivery problem, not a data-loss one.
Nothing has been lost
An application is stored on your site first, queued second, and delivered third. The candidate is only told it worked once it is safely stored, so every application anybody has submitted is still there waiting. That is worth saying to whoever is panicking before you start fixing anything.
First
Find out what is waiting
wp w3rp application-outbox status
wp w3rp application-outbox list-unresolvedThe <code>w3rp_delivery_backlog</code> health check reports the same thing in Tools → Site Health, so a queue that has stopped moving tells you rather than waiting to be noticed.
Causes
The three causes
In the order to check them.
- The Bullhorn connection is broken. Deliveries cannot complete without it. Check w3rp_connection, fix it, and the queue drains on its own.
- The queue is not being processed. If WordPress’s scheduler is disabled on your hosting, nothing in the queue moves. Same cause as the sync not running, and the same fix.
- Individual applications need a decision. Some are deliberately held rather than delivered — most often where the candidate may already exist in Bullhorn. Those need a human, and the next article covers them.
Fix
Clearing the queue
# Process the background queue, where the scheduler is not doing it.
wp w3rp queue run
# Move stored applications into the delivery queue.
wp w3rp application-outbox sweep
# Deliver what is queued, now.
wp w3rp application-outbox run
# Confirm.
wp w3rp application-outbox statusFailed deliveries are retried on a backoff rather than abandoned, so a brief Bullhorn outage resolves itself without anyone intervening. It is a queue that has stopped moving for hours, not one with a few retries in it, that needs attention.
Do not replay blindly to clear a backlog
If applications are sitting in an unresolved state, find out which state before re-queuing them. Replaying an application whose delivery outcome was never established can create a duplicate candidate record in Bullhorn, which is a worse problem than the one you started with.