Over engineering is a common pitfall for technical founders starting their own business. Maybe there's new hype about "framework X" or we definitely want to learn "technology Y" but this is often fraught with pitfalls and time wasting.

Realistically if we have existing knowledge and skills to be able to deliver a project, we should use them. Users do not care what technology is being used and as our default position is failure we need to be delivering early, learning and evolving.

Of course should one want to combine both creating a project with learning something new, this can be a great way to learn, however for Kukri Labs projects this is predominantly not the case. Technology can, in certain situations, also be a key differentiator so we remain open to using the right tool for the job but the majority of work will be done with familiar, mature tech.

Approved list

From the beginning, we define an approved list of technologies to take time out of making these decisions in the future. While this is a living document the idea is that if it's not on the list, don't choose it (unless there's a very good reason to research something new).

Frontend

Name	What is it?	Experience	Notes
Angular	SPA & SSR (with Angular Universal) framework	Extensive	Typescript native. Opinionated. Lots of out-of-the-box functionality.

Note: For API driven applications, Angular is the preferred SPA. For server rendered applications, the relevant templating language will be used.

Backend

Name	What is it?	Experience	Notes
Django	Python web framework	Extensive	Lots of out-of-the-box functionality
Pocketbase	'Backend as a Service' written in Go with SQLite DB	Go - extensive; Pocketbase - Medium	For simple, lightweight APIs

Data storage

Name	What is it?	Experience	Notes
SQLite	Relational DB	Extensive	Fast. Suitable for a lot of read-heavy projects. Has limitations but unlikely to reach with small projects. Is simple.
PostgreSQL	Relational DB	Medium-Extensive	Battle tested database. More complex but good if needing something more advanced than SQLite
Redis	In memory database and cache	Medium-Extensive	Battle tested database. Fast reads and writes and can persist to disk if necessary.
S3 (or compatible)	Blob storage	Extensive	Any S3 compatible storage for blob data. Likely Cloudflare R2.

Infrastructure

Name	What is it?	Experience	Notes
Hetzner servers	Cloud and dedicated servers	Extensive	Low cost servers. CAX{xx} Arm cloud servers; EX{xx} dedicated servers
Debian	Linux operating system	Extensive	Simple, performant and well documented
Ansible	Configuration management	Extensive	Repeatable configuration of servers. Used for quickly bringing up new servers as needed
Podman/Docker	Containers for applications	Extensive	Immutable infrastructure to build application and then promote. Easy to isolate different applications on same VM this way.

Note: We have chosen to explicitly avoid Kubernetes at this time as it is just something else to manage. Podman containers can be isolated using separate networks and run by Systemd so that should be good enough

Other

Name	What is it?	Experience	Notes
Ghost	Blogging and content creation platform	Medium	Opinionated and restricting at times however has email management, analytics etc. built right in with first class support. Work with the project.
Cloudflare	CDN; Serverless compute; Edge K/V database	Medium	DDoS protection for projects.
Tailscale	Wireguard VPN service	Medium	Secure connections to & between servers.
Material design	Google design guidelines	Medium	Rather than reinvent designs for common elements, use an off-the-shelf design system that integrates well with Angular & Django, is accessible etc.