Skip to main content
An alert is a hypothesis, not a fact. Ordyn treats it that way. Before anything posts to Slack, Ordyn checks whether the alert reflects real impact, what caused it, and whether it is the same problem you already know about.

​
What triage does with each alert

  1. Investigates. Ordyn pulls the relevant logs, metrics, traces, and recent code changes and works out what actually happened.
  2. Grades it. Each alert gets a severity: Low, Medium, High, or Critical. The grade reflects observed impact, not the monitor’s configured priority.
  3. Groups it. Alerts that share a cause are folded into one episode, even across different monitors. An episode is one condition going bad and coming back.
  4. Decides whether to post. The channel’s sensitivity level sets the floor. Below it, the alert is held. Critical always posts.

​
What this looks like in Slack

  • One message per episode. When more alerts join the same problem, the message is edited in place rather than posting again.
  • Recoveries update the same message. When the condition clears, the card says so. Ordyn does not post a separate “resolved” message.
  • Repeats do not re-page you unless the problem is materially different, has gotten worse, or a recorded fix did not hold.

​
Which alerts get triaged

Alerts from monitors Ordyn created are triaged automatically. Ordyn can also triage the alerts you already have. Turn on Auto-investigate for a Slack channel, a Linear team, a Jira project, or an Asana project, and Ordyn investigates new items there and replies in the thread. See Alert sources. Held alerts are not lost. They stay on the Episodes page as evidence, count toward recurrence, and feed into open issues.