I've been implementing AuditBoard for a client over the last quarter, and we've hit a critical roadblock. The approval workflow for control testing is stuck in an infinite loop between two approvers. The status shows "Pending Approval," but any action from either user simply triggers a notification to the other, with no progression to "Approved."
Support's response has been to check our "Workflow Settings" against their documentation, which we've done exhaustively. Their next suggestion was to clear browser caches and use incognito mode, which is irrelevant for a systemic workflow issue. We've been passed between tier one and tier two for over a week without a concrete escalation path.
This isn't just a bug; it's a process failure that halts the entire audit cycle. Has anyone else encountered a similar loop in the workflow engine, particularly after the recent platform update? I need to understand:
* If there's a known issue with conditional approval paths.
* Whether the root cause is in the workflow definition itself or the underlying queue processing.
* How you finally resolved it, especially if it required a support ticket to be escalated to engineering.
The client's tolerance is wearing thin, and we're now evaluating the operational risk this introduces to our timeline. A platform that can't reliably route approvals is a serious liability.
Trust but verify — especially the fine print.
Ah, the classic "clear your cache" deflection for a server-side logic loop. Classic.
You won't like this, but in my experience with these platforms, the "infinite loop" is almost always a configuration problem they've engineered themselves. Their workflow designer probably has a hidden state or a conditional you missed that's creating a circular reference. The recent update likely just exposed it.
The real issue is their support path. A week bouncing between tiers for a process-halting bug is unacceptable. You need to stop asking about the root cause and start demanding their escalation protocol in writing. Ask for the ticket to be flagged as a "business-critical outage" - that phrase sometimes trips a different SLA.
Have you checked if both approvers have *identical* permissions? Sometimes a "review" role versus an "approve" role, even if named the same, can cause this ping-pong.
If it's free, you're the product. If it's expensive, you're still the product.
You're absolutely right about the "clear your cache" deflection for a server-side issue. I've seen this pattern before, and your point about *identical* permissions is crucial. It often goes beyond roles to the specific workflow step permissions. I'd add that you should check for any "auto-reassign on reject" logic that might be misconfigured, creating the loop independent of user action.
Regarding escalation, insisting on the "business-critical outage" flag is solid advice. In my experience, you also need to find the internal name for their severity matrix. Ask them directly, "What is your Severity 1 definition and mean time to acknowledge?" That forces them to reference a specific policy document.
SQL is not dead.