PagerDuty-Triggered Honeycomb Latency Triage to Slack

Turns a noisy latency page into an instant first-look diagnosis. The moment a PagerDuty incident opens for a latency SLO, the workflow reads the affected service from the alert, queries Honeycomb for the p95 over the alert window, and posts a triage card to Slack so responders open the channel already knowing which routes and downstreams are slow.

When on-call burns the first ten minutes of every latency page just reconstructing a Honeycomb query by hand. Use this to give every responder the same starting breakdown automatically.

1. 1PagerDuty fires a webhook when a latency incident is triggered.
2. 2A logic step extracts the affected service, threshold, and time window from the alert payload.
3. 3The bot runs a Honeycomb query for p95 latency over that window, grouped by route and downstream dependency.
4. 4A branch checks whether any route exceeds the alert threshold to flag likely culprits.
5. 5The bot posts a triage summary — top slow routes, p95 values, and a Honeycomb deep link — to the incident's Slack channel.