TechPrufer
Sign inGet started

Blogs›Instacart›On-call Reasoning Harness

On-call Reasoning Harness

On-call Reasoning Harness

1
posts
2026

Introduced Blueberry, a Slack-native on-call reasoning harness designed to shorten time to first insight (TTFI) and time to test theories (TTTT) for on-call engineers. Blueberry automatically triages alerts within Slack threads, providing grounded explanations by correlating historical data (postmortems, runbooks) with live operational signals and changes. It features a durable, tool-aware architecture with a layered Model Context Protocol (MCP) for shared and team-specific investigators, enabling parallel evidence gathering and focused sub-agent investigations. The system captures reasoning traces and session context for inspectability and future improvement, aiming to transform tribal knowledge into reusable infrastructure and accelerate incident response.

2026

Blueberry: Force Multiplier For The On-Call Engineer

7/14/2026

Designed and implemented Blueberry, a Slack-native on-call reasoning harness. Key technical contributions include: a durable, Postgres-backed job management layer for enqueuing diagnostic passes; a processor for resolving team context and layering team-specific skills onto a shared baseline; a three-tiered Model Context Protocol (MCP) surface (in-process, shared server, team-hosted) for accessing diverse data sources; a dynamic agent loop capable of fanning out into parallel sub-agents for evidence collection; and persistence mechanisms for capturing session context and reasoning traces. Achieved an average TTFI of 3 minutes and TTTT of 3 minutes for auto-triggered triages and engineer-initiated investigations, respectively.