Hey folks. Been wrestling with a Zendesk setup for our cloud support teams and wanted to see how others are solving this. We have a central "Cloud Alerts" queue, but high-severity tickets (think P1 incidents, security issues) need to instantly route to the correct on-call team—SRE, Cloud Security, or Cost Engineering—based on the alert type.
Our current setup uses a combination of tags and basic triggers, but we're seeing delays and misrouting during off-hours, which is obviously not great. I'm worried our logic is too brittle.
I know Zendesk has features like Skills-Based Routing, Groups, and good ol' fashioned automation. For those of you handling critical, time-sensitive tickets, what's your playbook?
Here's a simplified version of a trigger we're testing, but it feels clunky:
```json
// Trigger: Route High Priority Cloud Alert
Conditions:
- Ticket tags: Contains "p1_alert"
- Ticket form: is "Cloud Incident"
Actions:
- If tag contains "security_vulnerability", set group to "Cloud Security Team".
- Else if tag contains "service_outage", set group to "SRE On-Call".
- Else if tag contains "cost_anomaly", set group to "Cost Engineering".
- Add tag "high_priority_routed".
- Set priority to "Urgent".
```
Is there a more robust way? Maybe using ticket fields to dynamically assign a group, or even integrating with something like PagerDuty for the actual on-call handoff? I'm also curious about escalation paths if the first group doesn't respond within a set SLA.
Love to hear what's working in your shops. The goal is zero human triage for these critical tickets.
cost first, then scale
Your trigger's logic is backwards. You're checking for tags *after* the trigger fires, which adds a step and can fail. Build the logic directly into the trigger conditions.
We moved off tag-based triggers for this exact reason. Our setup now uses the ticket's custom fields (like "Alert Type" dropdown) as the primary condition. Triggers fire instantly on ticket create/update.
Example trigger condition set:
* Ticket form is "Cloud Incident"
* Priority is "Urgent"
* Custom field "Alert Type" is one of ["Security", "Service Outage", "Cost Anomaly"]
Then the action is a single, direct group assignment. No nested if/else in the actions list. Much more reliable. The tag is just for auditing.
Benchmarks don't lie.
Yeah, that trigger is a Rube Goldberg machine for tickets. Off-hours delays scream "conditional actions" to me.
You're checking tags *inside* the actions list, which means the trigger fires, then it has to evaluate that if/else chain. That's your lag and your brittleness. The routing logic should live entirely in the trigger's *conditions*.
We split ours into dedicated, single-purpose triggers:
* Condition: Form "Cloud Incident" AND Priority "Urgent" AND Custom field "Alert Type" = "Security"
* Action: Assign to "Cloud Security Team", set priority to "Urgent", add tag "p1_security_routed"
Do that for each team. It's more triggers but they're atomic and fail fast. The custom field dropdown is key - it forces the alert source to classify it upfront, so the logic doesn't depend on tag naming conventions that someone might fat-finger at 3 AM.
Also, for cost anomaly alerts, we have a second automation that fires after 5 minutes if the ticket isn't touched, to ping our Slack war-room. Because let's be honest, nobody watches the queue at 2 AM unless money's actively burning.