Cutting database metadata load times by 80% and adding interactive ER diagrams

Ehitel Rodriguez — Mon, 24 Aug 2026 15:25:28 +0000

When working with relational databases, slow metadata fetching and lacking visual schema tools quickly become friction points in daily workflows.

In the latest update of EZQL, I focused on solving two main problems: eliminating connection bottlenecks when inspecting database schemas and introducing an interactive diagram tool.

Parallel Metadata Engine in Go

Expanding a database in an object explorer requires querying multiple system catalog views: tables, views, stored procedures, functions, triggers, users, and roles.

Doing this sequentially over a single connection took around 2.0 seconds on moderate schemas. To fix this, I reworked the backend service in Go:

Queries for all 7 object categories now run concurrently across the connection pool using sync.WaitGroup and goroutines.
Response time dropped from ~2.0s down to ~200ms (>80% reduction).
Added a 150ms debounced prefetching mechanism for column metadata to keep IntelliSense responsive in the editor.

Interactive Entity-Relationship Diagrams (ERD)

Reviewing foreign key structures directly from table trees can be tedious, so I added a dedicated visual diagram canvas:

Draggable table cards connected by dynamic Bezier curves based on foreign key relationships.
State persistence using localStorage, keeping node positions intact across sessions.
Export options for SVG, PNG, and clipboard copy.

Trigger Management and Object Tree Cleanup

Direct status detection for table triggers using sys.triggers (is_disabled).
Context menu options to enable, disable, alter, or drop triggers without manually writing the DDL statements.
Visual distinction between Primary Keys, Foreign Keys, and column data types.

If you work with SQL Server on macOS, Linux, or Windows, you can test these additions directly on ezql.app.

Do you usually rely on visual diagrams when working with existing schemas, or do you prefer inspecting constraints via queries?

Moving to macOS broke my SQL Server workflow. So I'm building my own client.

Ehitel Rodriguez — Sat, 22 Aug 2026 23:56:33 +0000

When I switched my main workstation to macOS, I hit an unexpected wall: managing Microsoft SQL Server outside of Windows.

For years, SQL Server Management Studio (SSMS) was my daily driver. Once on a Mac, the reality set in quickly:

SSMS is strictly Windows-only.
Azure Data Studio was discontinued.
Available alternatives either hid behind steep paywalls or lacked the essential everyday workflows I relied on.

That gap pushed me to start building EZQL.

The Vision

I didn't just want a lightweight data viewer. My goal is to build a unified, cross-platform client (macOS, Linux, and Windows) that steadily works toward the depth and feature parity of SSMS—while keeping the interface modern, focused, and intuitive.

Windows users aren't left behind either; the idea is to provide a consistent, native-feeling experience across any OS without unnecessary bloat.

The Ongoing Journey

Building a database client from scratch is full of interesting engineering challenges:
Streamlining query execution flows.
Efficient schema handling and state persistence.
Building clean, responsive result grids that stay out of the way.

Let's Discuss

To anyone working with SQL Server across macOS, Linux, or Windows:

What are the non-negotiable SSMS features or workflows you rely on the most every day?

DEV Community: Ehitel Rodriguez

Cutting database metadata load times by 80% and adding interactive ER diagrams

Moving to macOS broke my SQL Server workflow. So I'm building my own client.

The Vision

The Ongoing Journey

Let's Discuss