Data & AI
Database services: design, optimization and administration
The database is the one part of your system you cannot simply redeploy after a mistake: it holds the real state of your business. We design, optimize and administer PostgreSQL, MySQL, MongoDB and Redis databases for business applications, SaaS products and e-commerce sites.
This service is for teams running a database in production who face slowdowns, incidents or doubts about their backups, as well as new projects that want a solid schema from day one. We work on the data model, indexes, queries, server configuration, connection pooling and the backup and restore strategy.
Our approach always starts with measurements: execution plans with EXPLAIN ANALYZE, index usage statistics, slow query logs. We change nothing until we understand why the database behaves the way it does, and every change is tested on a copy before it reaches production.
When do we step in?
Queries getting slower over time
Your application slows down as tables grow. Pages that took a few milliseconds now take several seconds, usually because of missing indexes, N+1 queries or poorly written joins.
Schema designed in a hurry
The data model was created quickly at launch and never revisited. Catch-all columns, no foreign keys, wrong types: every new feature becomes harder to build.
Backups that were never tested
Backups may exist, but nobody has ever checked that they restore. Without regular restore tests, you know neither your real recovery point nor how long it takes to get back up.
Single database with no redundancy
One server hosts the production database. A hardware failure, a full disk or a failed update stops the whole business, with no replica to take over.
Choosing an engine for a new project
You are hesitating between PostgreSQL, MySQL, MongoDB or a combination with Redis. The right choice depends on the nature of the data, the volume, transactional needs and the team's skills, not on trends.
How we work
- 1
Audit and measurements
We collect the configuration, table sizes, index statistics, slow query logs and execution plans. The goal is to measure before touching anything.
- 2
Schema modeling or redesign
We design or fix the data model: appropriate normalization, precise types, integrity constraints, foreign keys, partitioning when the volume justifies it. Schema changes are written as versioned migrations.
- 3
Query and index optimization
We rewrite expensive queries, add or drop indexes based on their actual usage, and tune the server configuration: memory, cache, autovacuum, connection pool.
- 4
Backups and high availability
We set up automated, encrypted and tested backups with a suitable retention policy, then, depending on your needs, replication and a failover mechanism to a replica.
- 5
Monitoring and follow-up
We install dashboards and alerts on the indicators that matter: query latency, active connections, locks, replication lag, disk space. Regular follow-up prevents rediscovering the same problems six months later.
Technologies we use
- PostgreSQL
- MySQL / MariaDB
- MongoDB
- Redis
- Amazon RDS and Aurora
- PgBouncer
- pg_dump and pgBackRest
- Percona Toolkit
- Prometheus and Grafana
- Flyway / Liquibase
Related expertise
Why choose Agencei?
Decisions based on measurements
Every recommendation is backed by execution plans, real statistics and tests on a copy of your data, not by generic rules.
Application and infrastructure perspective
We also build the applications that use these databases. We know that a database problem is often fixed in the code, the ORM or the connection pool, not only in the server configuration.
Reversible changes
Versioned schema migrations, a backup before every intervention, a written rollback plan: we never touch a production database without a safety net.
Knowledge transfer to your team
We document what we do and explain the choices to your developers, so the database stays maintainable without depending on us.
In brief
- What is this service?
- Design, modeling, optimization and administration service for PostgreSQL, MySQL, MongoDB and Redis databases, including backups and high availability.
- Who is it for?
- Businesses, SaaS vendors and e-commerce sites whose application relies on a production database, or who are starting a project and want a solid schema.
- What problem does it solve?
- Slowness caused by indexes and queries, poorly designed schemas, untested backups, lack of redundancy and uncertain engine choice.
- How long does it usually take?
- An audit usually takes a few days. A targeted optimization takes a few days to a few weeks. A schema redesign or a high availability setup is often measured in weeks.
- What factors influence the price?
- Price depends on data volume, the number of engines and environments, the state of the existing schema, availability requirements and whether ongoing follow-up is included.
- How does the engagement run?
- Audit with real measurements, prioritized action plan, changes tested on a copy, deployment through versioned migrations, then monitoring and follow-up.
- What are the risks?
- Working without prior measurements or backups can degrade performance or lose data. Schema changes must be coordinated with the application code.
- What alternatives exist?
- Staying on a managed service without optimization, training the in-house team in administration, or hiring a full-time database administrator if the volume justifies it.
Frequently asked questions
Should I choose PostgreSQL, MySQL or MongoDB?
PostgreSQL is our default choice for relational data and business applications, thanks to its rigor and its features (JSONB, window functions, extensions). MySQL remains relevant for ecosystems that already rely on it, such as WordPress or PrestaShop. MongoDB makes sense for documents with variable structure or very high write volumes. Redis usually complements these engines as a cache or queue, rarely as the primary database.
Can you work on a database hosted by a cloud provider?
Yes. We work on Amazon RDS, Aurora, MongoDB Atlas and self-hosted instances. Managed services simplify backups and replication, but the schema, indexes and queries remain your responsibility, and that is where most of the gains are.
How do I know if my backups are reliable?
A backup is only reliable if it has been successfully restored recently. We set up automated restore tests on a separate environment, verify the integrity of the restored data and measure the time it takes. Both values, recovery point and restore duration, must be known before the incident.
Does optimization require downtime?
Rarely. Most indexes can be created online (CREATE INDEX CONCURRENTLY on PostgreSQL), query rewrites ship with the application, and configuration changes often only need a reload. Genuinely blocking operations are identified in advance and scheduled during off-peak hours.
Do you offer ongoing administration?
Yes, as regular follow-up: monitoring, slow query review, minor updates, verification of backups and replicas. The frequency is adapted to how critical your database is and to the size of your team.
Related services
PostgreSQL expertise
Schema design, optimization, replication, version upgrades and migration to PostgreSQL.
See this serviceMongoDB expertise
Document modeling, indexes, aggregations, migration and performance on MongoDB and Atlas.
See this serviceData migration
Migration between databases, engines, versions or systems, with ETL, validation and zero-downtime cutover.
See this servicePerformance Optimization
Profiling-based diagnosis and targeted fixes: Core Web Vitals, SQL queries, caching, CDN.
See this serviceAPI and backend
REST and GraphQL APIs, Node.js, Spring Boot and Python backend services, documented and secured.
See this serviceCommon problems
Tell us about your project
Describe your need in a few lines: we come back to you with a first analysis and the next steps.