Automatically generate incident names and summaries from alerts

Tired of waking up to incident names like: [CRITICAL] k8s-prod-cluster-east1: High pod eviction rate (>45%) detected in default namespace causing cascading OOM errors across statefulset replicas (oops-i-oomed-again-service-v2)?

Would you prefer it said:

• Name: “High Pod Eviction - Memory OOM in Production StatefulSet”
• Summary:
  • Widespread pod evictions occurring in the east1 production cluster.
  • Over 45% of pods are being OOM killed.
  • oops-i-oomed-again-service-v2 affected, with failures spreading across replicas.

We've just rolled out AI-powered incident naming and summarization that automatically generates clear, descriptive names and summaries based on your alert data.

You can enable AI naming for specific incidents in your alert route while keeping manual naming for others, giving you the flexibility to customize based on your team's needs.

For example, you might want to:

• Automatically name incidents from customer support tickets based on their content.
• Generate descriptive summaries for monitoring alerts that include relevant metrics.

Also, we don’t modify the name of the underlying alert - meaning that if you, like us, have gotten used to blearily reading alert-name-soup, you’ll still be able to make use of your pattern-recognition abilities!

Grafana integration upgrades

When alerts come in from Grafana, the first step is often to check the dashboard that triggered the alert. Now, with upgrades to our Grafana integration, we’ll pull a screenshot of the dashboard and place it in the channel.

You can enable this by visiting Settings → Integrations → Grafana.

🚀 What else we’ve shipped