Zendesk SLA Watch

Watch your open Zendesk tickets and push the current list to your team each run, so an agent can spot and act on tickets approaching an SLA target before the clock runs out.

Use case

Support teams commit to SLA targets (first reply, next reply, resolution). A next-day report is too late — someone needs visibility into the open queue now. This workflow polls Zendesk on a schedule, fetches the open tickets, and delivers them to your subscribers so an agent can triage anything at risk.

What it does each run

1. fetch-tickets (zendesk) — calls the Zendesk REST API with operation: list_tickets. Like every InTouch connector, it publishes exactly two outputs: - {{fetch-tickets.result}} — the raw JSON API response (string) - {{fetch-tickets.status}} — the HTTP status code
2. notify (message) — sends a notification to the workflow's contacts whose body embeds {{fetch-tickets.result}} (the ticket payload) and reports the API status {{fetch-tickets.status}}.

If the Zendesk step fails (auth, rate limit, network), alertOnError notifies the intouch user so the watch never fails silently.

Required credentials

• Zendesk API key — the apiKey property is set to the placeholder <set via credential>. Supply the real value yourself, either by attaching a Zendesk credential to the task or via a runtime environment variable; never hardcode the key in the workflow. Create the token in Zendesk Admin Center → Apps and integrations → APIs → Zendesk API. The Zendesk API token is a static secret — no OAuth round-trip needed.

Setup

1. Supply the Zendesk apiKey via credential or runtime env (see above).
2. Subscribers — link at least one contact/alert to this workflow (email, Slack, SMS, etc.) so the message step has somewhere to deliver.
3. Schedule — attach a schedule (e.g. every 15 minutes during business hours) so the watch runs continuously.

Customizing

To produce a human-readable digest instead of raw JSON, insert an ai (or runtimeenv script) task between the two steps that consumes {{fetch-tickets.result}} and emits a summary, then reference that step's output in the message body.