Granular is database performance software from TechSQL. It finds what's slowing MySQL and PostgreSQL down, tells you exactly how to fix it, and proves the fix worked.

Read-only collectorMySQL 8 · PostgreSQL 14–17Hosted or self-hosted
Live demo

Three incidents, start to finish

A simulation of the Granular console running in your browser: a missing index after a deploy, a lock pile-up behind one forgotten transaction, and partitions about to run out. Switch scenarios below. All data is synthetic.

Granular · Missing indexsimulated data
shop-orders-prodMySQLHealthy
0.92
Active sessions
2.4 ms
Per statement
0/5
Steps done
Database time (average active sessions)last 30 s
0.024usual for this hour
Watch
Everything normal. Granular samples active sessions every second and learns what's usual for each hour.
Products

The Granular family

Database tooling built around one idea: a recommendation isn't done until it's verified and its value is measured.

Pilots open

Granular for MySQL & PostgreSQL

Continuous monitoring, diagnosis and tuning for MySQL 8 and PostgreSQL, from one console.

  • Wait-time analysis, live sessions and blocking chains
  • Index, query, configuration, schema and partitioning findings with fix and rollback SQL
  • Plan-change detection and before/after verification
  • Health reports and savings estimates for management
Start a pilot →
Databricks

Granular for Databricks

The first Granular product: performance tuning for Databricks workloads, with the same detect, fix and verify approach.

Ask about Databricks →
Services

Performance consulting

Database performance reviews and tuning engagements from the team that builds Granular, using the same method and evidence.

Discuss an engagement →
How it works

From a slow query to a measured saving

A small collector runs next to your database, reads its built-in statistics views once a minute, and sends redacted snapshots over HTTPS.

  1. 01

    Detect

    DB time, waits and latency against what's usual for that hour and weekday. Alerts only when it's sustained.

  2. 02

    Diagnose

    The statements, waits, blocking chains and plan changes behind the slowdown, in one place.

  3. 03

    Recommend

    A specific fix with evidence, the SQL to apply, the rollback, and its expected effect.

  4. 04

    Verify

    Before and after at the same traffic, with an SLA check and a scan for side effects.

  5. 05

    Measure savings

    Freed database time translated into vCPU-hours and money, labelled as an estimate.

Why Granular

Built for the people who sign off on changes

Proof, not a dashboard

Each change is judged on per-call latency at the same traffic, so a quiet afternoon can't pass for an improvement.

Safe on production

Read-only statistics access. Plans are computed without running your queries. Nothing is applied without you.

Private by construction

Every literal value is stripped from SQL text before it leaves your network. Table data is never read.

Two engines, one console

MySQL and PostgreSQL get the same depth: waits, plans, schema, partitioning and configuration.

Your choice of hosting

Use the hosted service, or run the self-hosted edition inside your own network with a license file.

Numbers finance understands

Monthly health reports and savings estimates show what tuning was worth, clearly labelled as estimates.

Company

TechSQL Consulting Services

TechSQL builds database performance tooling under the Granular name. Granular for Databricks came first; Granular for MySQL and PostgreSQL brings the same approach to the operational databases most businesses run on.

Our principle is simple: every recommendation carries its evidence, the exact change, how to roll it back, and the effect you should expect. Then we measure whether it delivered.

Talk to us

Pilots, the self-hosted edition, Databricks, or a performance review: write to us and we'll get back to you.

hello@techsql.com
Email usSign in to Granular