Data
PostgreSQL
Primary relational database for new products.
How we use PostgreSQL
Postgres is our default for transactional data, search extensions, and JSON where it belongs.
Where we reach for it
Transactional core data
Orders, accounts, bookings and anything where constraints, foreign keys and ACID transactions protect data integrity.
Mixed relational and JSON
Structured tables with a JSONB column for flexible attributes, indexed and queryable without a second database.
Search and geodata
Built-in full-text search for moderate needs, and PostGIS when locations and distances are part of the product.
Reporting on live data
Window functions, CTEs and materialised views for dashboards, so simple analytics do not need a separate warehouse.
When we pick something else
For very high-volume search relevance or log analytics we add Elasticsearch next to it. For pure caching, queues or ephemeral state, Redis fits better. If an existing system runs well on MySQL or MariaDB, we leave it there unless there is a reason to move.
What we watch for
Backups and restores
Continuous archiving or regular dumps are only worth something if restores are tested. We rehearse recovery, not just schedule backups.
Vacuum and bloat
Heavy update workloads need autovacuum tuned and monitored, or tables and indexes grow and slow down.
Connection management
Each connection costs memory. A pooler such as PgBouncer keeps app instances from exhausting the server.
Common questions
- PostgreSQL or MySQL for a new project?
For new products we default to PostgreSQL: stricter SQL, richer types, JSONB, extensions. MySQL remains a sound choice where the ecosystem or hosting already depends on it.
- Can PostgreSQL be hosted in Switzerland?
Yes. It can run on a Swiss VPS, on-premises, or with Swiss and EU providers offering managed Postgres. We pick the setup with your data residency requirements under nDSG and GDPR in mind.
- Can you migrate our MySQL database to PostgreSQL?
Yes. We map schema and types, migrate data in rehearsed runs, adjust queries that rely on MySQL behaviour and verify row counts and checksums before cut-over.
Next step
What is costing you time, money or customers today?
You get an honest assessment of whether and how we can help, followed by a written plan with a cost estimate and timeline.
Tell us what you're buildingWe reply within 1–3 working days, by email, call or video. Your choice.