Learn depth and internals of software, engineering of scale and have fun doing it!
Delve into the code, architecture, & innovations shaping our digital world.
Quotes from article on how #postgres indices work:
> Every table in postgres has one or more corresponding files on disk, depending on its size. This set of files is called a heap and it is divided into 8kb pages.
#database
Multi-column indexes are generally going to be smaller and faster than using multiple indexes, but they’ll also be less flexible. The database can search for a subset of the indexed columns, as long as they are the leftmost columns.
> If reading from registers is like getting data my head, then reading from disk is like getting data from pluto.
- CS186Berkeley
https://t.co/2HPSscu4Mx
#harddisk#storage#CS186Berkeley
“Processes are entirely operating system abstractions, not something CPUs natively understand or keep track of.”
- Amazing human who built https://t.co/K7GfXC4EdW
Domain knowledge - e.g. operator can monitor its application and react to errors with specific behavior to resolve the error instead of default behavior of restarting the containers.
#kubernetes#operators#k8s#whitepaper
Went though the Kubernetes Operator whitepaper, here are some highlights 🧵
The core purpose of any operator is to extend its orchestrator's underlying API with new domain knowledge.
Operational automation - Operators can extend and enhance orchestrators with higher-level automation such as deploying clustered software, providing automated backups and restores, or dynamic scaling based on load.