"On our t3.small (2GB RAM), the web container uses ~780MB. Spawning an exec container for a quick database query adds another ~500MB. Two concurrent exec calls during a deploy? That's 2.5GB on a 2GB machine. OOM killer takes the web process down."
I suspect this is the reason for SQLite and not PG or some other DB. From my read of the article, it sounds like they're running lean (single host for web, db, etc). PG et. al. add overhead that they likely can't (or more likely won't want to) afford on that box. Maybe I'm wrong?
Generally, it's machine generated. I've only ever typed it out manually if I needed something specific for a unit test and that only for small messages. It's tedious to do by hand.
A shop I worked in ages ago had a mainframe on site supporting lots of batched transactions over the course of a 24 hour period, along with the daily operations for a 35 facility logistics company, spread throughout the southeastern United States. They had a staff of 8 mainframers that ran the entire operation. All of the PCs onsite (my area of responsibility) were basically 3270 terminal clients for z/OS.
Sometime later, a connection on LinkedIn pointed out the following resources:
Also, mainframes are really good at high throughput transactional jobs. That's why you see them in banks, transportation, insurance, etc. Big Tech™ doesn't see it as "cool" and are too focused on the Next Big Thing™, so there's not a lot of attention there. Sometimes, boring just gets the job done.
This was essentially my experience with WGU as well. I started in November and was done by the following August. Most of my classes from community college were transferred and I had ~90 CUs (EDIT: Just a guess on the number. It's been a while...) I to complete. I was able to complete several courses a week for the first few months, but started tapering off to about 1 course a week as time went on. I was exhausted by the end, but it was worth every penny and minute spent.
I suspect this is the reason for SQLite and not PG or some other DB. From my read of the article, it sounds like they're running lean (single host for web, db, etc). PG et. al. add overhead that they likely can't (or more likely won't want to) afford on that box. Maybe I'm wrong?