Fast queries, backups that restore, migrations without surprises
We find the queries that slow your application down, prove that your backups restore, and plan migrations so that the cutover is a rehearsed step.
Before1840 ms
Seq Scan on orders 4.8 M rows
After12 ms
Index Scan using orders_customer_created_idx- First call
- Free, no commitment
- After the call
- A written proposal: scope, steps, price
- Price
- Agreed in writing before work starts
The problem
Where it usually hurts
The application got slow as the data grew. Someone suspects the database, more hardware was added, and the bill went up while the slow pages stayed slow.
Backups run every night and the job reports success. Nobody has restored one in a long time, so nobody knows how long a restore takes or whether it works at all.
A version upgrade or a move to another platform has been on the list for a long time. It is postponed again and again, because nobody wants to own the cutover.
Scope
What you get
- A list of the slowest queries with their plans, and the fix for each: index, rewrite or schema change
- A restore drill: a backup restored to a separate host, timed and checked against the source
- Backup and restore procedures written down and scheduled, so the drill is repeated
- A migration plan with rehearsals, a cutover window and a tested rollback
- Monitoring for slow queries, replication lag, disk growth and failed backups
- A written report of what we changed and why, for your team and your auditors
Process
How an engagement runs
Four steps. You decide after each one whether there is a next one.
- 01
Book a call
Tell us what hurts: the release process, an audit finding, a slow database, a cloud account nobody fully understands. We ask questions and say plainly whether we can help.
- 02
Get a written proposal
Scope, steps and price in writing. You decide with the facts on the table.
- 03
Work in your systems
We work in your repositories, your pipelines and your cloud accounts, alongside your team. Our changes go through review like any other change.
- 04
Keep everything
Code, configuration and documentation stay in your repositories. Your team runs the result without us.
FAQ
Questions we get before the first call
Which databases do you work with?
PostgreSQL. If you run a different database, tell us in the first call. We say plainly whether we have the experience for it, and if we do not, you hear that before any work starts.
Do you need access to production data?
For performance work we need query statistics and execution plans, not customer records. Where a task needs real data, such as a restore drill, it runs inside your environment under your access rules.
Can you fix performance without changing the application?
Often a part of it: indexes, configuration and maintenance need no code change. Some queries have to be rewritten in the application, and then we deliver the change or work on it with your developers.
How risky is a migration?
A migration is risky when the cutover is the first time anyone runs it. We rehearse it against a copy, measure how long each step takes and test the way back, so the cutover follows a script that has already worked.
Do you operate our databases afterwards?
The goal is that your team operates them, with the procedures and the monitoring we leave behind. If you want continuing support, we agree that separately.
Talk to us about Databases
Describe where you stand today. We tell you plainly whether we can help and what a first step would look like.
The first call is free and without commitment.