Bedrock Python¶
Foundations for
serious Python
services.
Open-source libraries for the infrastructure behind your service. Runtime, clients, data and reliability — built for the same production stack.
Managing PostgreSQL partitions, one failure at a time
Two ways a partitioned table gets you out of bed, and the library that grew out of them: a plan you can read before it runs, ownership that never drops a table it did not make, the same API async and sync, a command line and a container image for teams with no Python, hooks from a YAML document, and one page for the AI assistant doing the wiring.
Latest from the blog
See all posts →Zero-dependency cores
An infrastructure library ends up in every service, and so does every dependency it declares. Measured: three of eight install exactly one distribution, and the FastAPI extra adds twenty-two.
How I start a production-grade Python library in 2026
One command, forty-one files, a green gate in three seconds. What each group of files decides, and the line of configuration that selected a rule family and then ignored it.
Migrating from pg_partman to application-managed partitions
What happens to the partitions that already exist? Measured against pg_partman 5.5: nothing. Both maintainers run side by side, and ownership is decided by bounds rather than names.
Put the ideas to work.
Pick a task. Find the right starting point.
Build your service.
Composable Python foundations for your runtime, data and reliability.
Clear the busywork.
Focused tools to review code, find answers and plan database changes.