[Engineering Log] Building Echo Dual Interface
Turning Echo from an MCP-only memory server into a dual-interface local tool with agent workflows, CLI inspection, shared services, and direct maintenance paths for humans.
- #go
- #retrospective
- #mcp
Turning Echo from an MCP-only memory server into a dual-interface local tool with agent workflows, CLI inspection, shared services, and direct maintenance paths for humans.
A mid-level developer retrospective on shifting Echo memory deletion from content matching to ID-based operations to ensure AI tools target database rows with no ambiguity.
Fixing DuckDB file locks in Echo by replacing the startup connection with on-demand database sessions that open, sync, query, and close during each analytics request.
Adding local FinOps and Knowledge ROI analytics to Echo so persistent memory can measure cost, hit rate, carbon estimates, and low-signal context before quality drifts.
Rebuilding a Linktree-style Vue project as a Go static site generator to reduce npm maintenance, simplify builds, and learn Go through a small real-world publishing tool.
Building an MCP server with Go and SQLite to give AI agents scoped, local-first memory, reduce repeated context setup, and keep project decisions available in CLI sessions.
Fixing Grafana dashboard provisioning by moving from brittle embedded JSON in values.yaml to standalone files, resolving 'invalid character' bugs once and for all.
Executing a risk-based, four-stage migration of the observability stack from Docker to Kubernetes to ensure data integrity and system stability.
Migrating from EOL Promtail to Grafana Alloy via a Strangler Fig pattern, bridging Kubernetes and Docker logs to ensure zero visibility loss during the platform transition.
Eliminating static .env files by centralizing secrets in OpenBao via a reusable internal/secrets module, decoupling credentials to streamline the Kubernetes migration phase.
Why I migrated a local Python CLI script to an Azure Serverless Function to eliminate virtual environment friction and gain practical hands-on cloud experience.
A deep dive into Systemd vs Cron: Automating an observability hub with self-healing timers, centralized logging, and dependency management. Read the full guide to learn.
An investigation into B-Trees, the library analogy, and why we don't index every column. Discover essential tips, full code examples, and practical insights in this guide.