Auto-merge incidents to dedup alerts? Other options for consolidating alerts?

We are on the Pro plan

I have a challenging problem (for me at least). We have to alert on router WAN ports that don’t fail when the circuit goes down (carrier equipment on site that does not support OAM so the router port stays up). Monitoring log files for routing events is not desirable as it causes way to many false alerts and we would still have to deal with multiple alerts. To resolve this, we run tunnels over the interface so if the far side of the carrier equipment goes down, the tunnel fails so we know that the circuit is down. We have several tunnels to different locations for redundancy so we can get up to 4 tunnel alerts at the same time when the circuit fails. The source of the alert is the same host and the only difference is the tunnel number (eg Tunnel 1, Tunnel 2, Tunnel 3 and Tunnel 4). I was reading on Merge and dedup but not finding an automated solution yet.

Here is what the goal would be:
If PD gets more than 1 tunnel alert for the same host, it merges these alerts into a single notification.

Bonuses if possible:
Only alert if all 4 tunnels fail (maybe within a span of time?)

Here is an example of the alerts:
[CRITICAL: ‘Tunnel10000 Tunnel to smar-sys3a-r1: ATT AVPN PseudoID None CID DHEC.12345 Status’ on ‘marblefalls-r1’]
[CRITICAL: ‘Tunnel20000 Tunnel to smar-sys3a-r1: ATT AVPN PseudoID None CID DHEC.12345 Status’ on ‘marblefalls-r1’]
[CRITICAL: ‘Tunnel30000 Tunnel to smar-sys3a-r1: ATT AVPN PseudoID None CID DHEC.12345 Status’ on ‘marblefalls-r1’]
[CRITICAL: ‘Tunnel40000 Tunnel to smar-sys3a-r1: ATT AVPN PseudoID None CID DHEC.12345 Status’ on ‘marblefalls-r1’]

You could look at using event orchestration rules to create the same dedup key on the incoming critical tunnel alerts to something that would be the same for all of that site’s tunnel alerts. This would deduplicate them all into one alert and incident.

Your bonus approach might require a higher subscription to Event Intelligence and/or using something external to PD like an AWS Lambda, our Process/Runbook Automation offering, etc.