Skip to Main Content

Share your product feedback

Status Future consideration
Categories Build
Created by Erik Baerresen
Created on Jun 12, 2026

Service health status page (or notification channel) for Elle / Aha! Builder

What is the challenge?
  • There is no status page, notification channel, or in-product indicator for Elle / Aha! Builder service health. When the backend or its upstream AI provider has an incident, builders only discover it through failures: conversation-killing activejob errors that terminate the chat and destroy its full context, or chats that hang indefinitely on a spinner with no error at all. Support confirmed (June 2026) that no health-visibility channel currently exists and that transient job failures terminating whole conversations is a known weakness.

What is the impact?
  • During June 9–11, 2026, recurring upstream-provider incidents halted all development on our Builder application for the better part of three days. Four separate conversations were killed mid-build (losing their accumulated context each time), additional time was lost to retries during windows we didn't know were incidents, and one morning was spent probing a degraded-but-not-erroring backend (a trivial question hung 40+ minutes). Without health visibility, every failure costs a support round-trip to distinguish "our problem" from "platform incident," and teams burn chat context and rework discovering outages empirically.

Describe your idea.
  • Provide visibility into Elle / Builder service health so teams can pause work during incidents instead of discovering them through failures. Any (or ideally all) of: (1) a public status page covering Elle / Builder backend and its AI-provider dependency, like status.aha.io but scoped to Builder/AI service health; (2) an in-product banner in the Builder editor when a degradation or incident is active ("Elle is currently experiencing delays — your chat may be affected"); (3) a subscribable notification channel (email/webhook/RSS) for incident start and resolution. Complementary hardening: when a transient backend/job failure does occur, fail the single response — not the entire conversation — so accumulated chat context survives the retry.

  • Attach files