alibasharat5

Thank you so much really appreciate the kind words... You are right.Context loss during incident response is the problem we’re aiming to solve deeply, especially for teams that live in Slack and don’t want to chase dashboards.

Let me try to answer your great questions:

1. Conflicting advice or noisy threads: Right now, RobinRelay summarizes past replies per alert occurrence (thread), then aggregates them across all past occurrences of the same (e.g., “CPU spike on api-server”). If conflicting advice exists (e.g., restart vs scale up), we preserve each cause separately and display them like:

“In the past, this alert was caused by: – autoscaler OOM restart (3x) – DB connection saturation (2x)” We’re working on tagging which ones were “confirmed fixes” vs just investigative noise and will open that to user feedback soon.

Alert variants: We normalize alerts our algorithm similar to finger prints for vector db. We’re exploring multiple ways like fuzzy hashing, LLM's, custom way to detect meaningful variants (e.g., same service, different root cause) and eventually flag them as “related but distinct”.

Trend detection (e.g. precursor alerts): Yes! This is on our roadmap. We want to show timelines like:

If you're seeing these pain points first-hand, I’d love to chat or hear how your team handles it today — this kind of feedback shapes the product every week.

Thanks again for the thoughtful comment

Thanks, Yes, Exactl! RobinRelay reads directly from the Slack alert messages (from monitoring tools, custom bots, etc.) and builds memory from the messages and thread discussions around them or even discussions in the channel itself (not in thread) we build Semantic matcher algorithm to connect message with the alert..

The goal was to avoid needing any API keys or deep integrations - just install, select the alert channel, and it starts working by parsing alert messages and mapping patterns.

You’re totally right though, there is richer metadata inside tools like Datadog. But we think end users have better context of known issues, but we're intentionally starting frictionless and will expand further to provide insights on newly seen incidents and use API to pull all the data.

Glad to hear you’ve felt the pain too... that’s exactly why I built this!

Thanks for checking it out! Tech-wise: RobinRelay is built using Slack Events API + Supabase for alert history and AI Agents. We wanted to avoid building another dashboard. So the product works entirely in Slackfrom setup to insights to summaries. Would love to hear how your teams manage alert noise and if memory around incident history would be useful to you.