telegraf is so easy to use and extend

Definitely… you can write custom scripts that Telegraf will run and write that data to Influx. For instance, I have one that writes the Gateway status information from pfSense so I can track and graph any internet downtime.

CPU/RAM/Disk/Network etc. get written to Influxdb via Telegraf, and visualized with Grafana.

Logging and errors go to Graylog stack (Mongodb, Opensearch, Graylog).

Unraid

because it’s much, much faster and easier to consume content via video

That totally depends on the content. Using your example, yes, a video of an explosion is going be much more efficient than a block of text about the same explosion. But for something like this, I find it MUCH slower to try to glean the relevant information from a video than from an article. An article can be skimmed easily so I only have to focus on the parts that I care about. Skimming a video, on the other hand, is a pain. Also, if the content is a step-by-step how-to, the video might be OK as long as I can follow along in real time. However, if I have to keep pausing and going back to rewatch a section, then an article is going to be easier to work with.

If you look at the markings on the baffle in the T320, it’s marked to indicate the second CPU as well as the second bank of RAM slots. I think it’s safe to say it’s identical.

Any SATA or SAS drive should work just fine. If you need some hot swap caddies, you can buy them fairly cheap on Amazon: https://www.amazon.com/s?k=dell+t320+drive+caddy

I’ve personally used the WorkDone brand caddies, and they are perfectly usable!

edit: I’m fairly certain it doesn’t support four 3.5 AND eight 2.5… the form factor supported depends on the backplane that’s installed. Also know that the backplane physically supports double that number of hard drives… you’d just need an HBA card with two internal ports. See this list for some options: (https://forums.serverbuilds.net/t/official-recommended-sas2-hba-internal-external/4581)

Is there not a native Nextcloud app for your tablet?

That doesn’t make any sense… the fact that it’s only used in part of the world makes it even more useful for the bot to define it.

Is there a reason why your bot doesn’t define CSAM?

I use a Graylog/Opensearch/Mongodb stack to log everything. I spent a good amount of time writing parsers for each source, but the benefit is that everything is normalized to make searching easier. I’m happy with it as a solution!

Given the choice between RAID but no backup, and backup but no RAID, it’s backup but no RAID by a mile.

I like my Venstar Explorer Minis. They have both a local API and cloud, both of which can be disabled independently (so you can have local ON and cloud OFF).

I really think there’s nothing better than Photostructure in terms of viewing and re-discovering your photos. It’s still a young product going through growing pains, but the things it does, it does well.