Teams error Unable to reach the app: "Unable to reach the app" - invoke response timed out

This is the error Teams shows the user, not your logs, when an invoke-style interaction - a message extension action, a dialog (task module) submit - gets no answer in time. The messaging-extension documentation is exact: "Your app has five seconds to respond to the invoke message," and "If the app doesn't respond within five seconds, the Teams client retries the request twice before it sends an error message Unable to reach the app. If the bot replies after the timeout, the response is ignored."

Three facts in that sentence shape the fix. The budget is five seconds, wall-clock, including your cold start. The client retries twice - so a slow handler runs up to three times, which is why half-finished side effects (three database rows, three tickets) accompany this error. And a late response is discarded, so finishing at second six is indistinguishable from crashing. Microsoft's documented pattern: "The app must defer any long-running actions after the bot replies to the invoke request. The long-running action results can be delivered as a message" - acknowledge inside the window, do the real work asynchronously, deliver the outcome as a normal or proactive message.

// budget: 5 seconds, retried twice, late responses ignored
{
  "name": "composeExtension/submitAction",
  "type": "invoke",
  "value": { "commandId": "createTicket", "data": { ... } }
}

Architect every invoke path as acknowledge-then-work from the start; retrofitting async into a synchronous handler mid-incident is miserable. Watch p95 handler latency against the five-second ceiling as a standing alert. The per-surface response windows, including this one, are tabulated in response-time limits, and endpoint reachability itself is the subject of the webhook debugging guide.