Glossary · Platform
Stateful alerting
Stateful alerting is an approach where every alert is a tracked object with its own lifecycle state — open, acknowledged, escalating, resolved — rather than a fire-and-forget message, so the system always knows what is still outstanding and what to do next.
In practice
The alert remembers where it is
When an excursion fires, IRIS creates a durable object for that alert and moves it through a lifecycle: open → escalating → acknowledged → resolved. At any instant the system can answer “is this still outstanding, which round are we on, and who is next?” — because the alert is that state. A fire-and-forget message cannot; a chain of automations can only reconstruct it, which is why those chains grow harder to operate reliably as sites and languages multiply. Stateful alerting is what makes reliable escalation and a trustworthy audit trail possible at all: the same object that decides the next step also records it.
Stateful alerting — frequently asked
Why does alerting need to be stateful?
Because a critical alert is a process, not an event. To escalate correctly you must know whether the current alert is still unacknowledged, which round it is on, who has already been tried, and when to give up or resolve. A stateless message blast cannot answer any of those questions — so escalation, deduplication and proof all become unreliable.
How is stateful alerting different from chaining automations (e.g. Make.com or Zapier)?
Automation tools fire steps when triggered but hold no single source of truth for an open alert. Engineers end up reconstructing state across multiple scenarios, which is hard to operate reliably as volume grows. Stateful alerting gives each alert one owned, durable state object that drives the whole lifecycle — which is exactly why IRIS replaced 26 Make.com scenarios for Seemoto with one engine.
How does IRIS implement state?
Each alert is a Cloudflare Durable Object pinned to EU jurisdiction. It holds the alert’s state, drives up to ~20 escalation rounds, and records every transition immutably — so the same object that decides “escalate next” also produces the audit trail.
See how IRIS uses a stateful engine to escalate until a human acknowledges.
How escalation worksAlerting with a memory
Start free with 1,000 credits — no credit card — and send a stateful, escalating alert today.