The marketing mixes metaphors, talking about gardening, growing, curating… all part of sustainable process that includes plants dying.

It also uses words like forever and permanent.

Having content live forever is at odds with metaphors of the natural world, where things naturally die.

How does it handle helm features that are not valid compose features? Silent failure or loud warning?

To explain how HAproxy and competing tools solve this:

Two servers are prepared to be the single reverse proxy, but one is active. They constantly communicate with a “heartbeat”. When the active one fails to send a heartbeat, the secondary executes the steps to become the active primary. When the primary’s heart starts beating again, it becomes active again.

So there can be a few seconds of downtime, but the failover is automatic.

Yes, especially if people use the “latest” tag, trusting whatever the container might be updated to do in the future.

I am using Navidrome and if it has significant bugs, I haven’t run into them yet.

Part of the app resides on the GitHub infrastructure, where GitHub stores, processes and displays results. So their costs are not zero.

But GitHub could take a “tax the rich” approach to pricing by charging enterprise customers more for self-hostingand leave it free for others.

A lot of open source is funded like that— most funding for a project comes from a very few companies and everything else uses it free or for very low donations or costs.

You have never had some family member experience a broken website that they needed to work but you were not around to fix it on the server side?

Because less than 1% of users would use it and your trusting the security of not one bit partner but thousands of ever-changing small partners.

Also, email is already federated.

I haven’t heard of that happening much outside of law enforcement raid.

Laptops, yeah. But stories of homes being broken into to steal servers?

When was the last time you saw a headline: “Thieves steal home lab”?

The encoding format of URLs is URL encoding, also known as percent-encoding. Content in the URL may be first encoding in some other format, like JSON or base64, and then encoded additionally using percent-encoding.

While there is a standard way to decode percent-encoding, websites are free to use base64 or JSON in URLs however they wish, so there’s not a one-size-fits-all way to decode them all. For example, the “/” character is valid in both percent-encoding and base64-encoding, so to know if it’s part of a base64-encoded blob or not, you might end up trying decoding several parts of the URL as base64 and checking if the result looks like URL-- essentially brute force.

A smarter way to do this might be to maintain a mapping between your favorite sites that you want to decode and what methods they use to encode links. Then a tool could efficiently directly decode the URLs embedded in these click trackers.

Lol. After professionally hosting email for 15 years I’m happy to let someone else handle it now.

About 90% of incoming mail will be spam and it will be your job to make sure you are doing good job of classifying it so you don’t get junk in your inbox and don’t lose real mail in the spam folder.

Then for outgoing mail you need to make sure SPF, DKIM and DMARC are all in order.

Then there is all the usual stuff of security updates, backups, monitoring, alerting, logging and having a plan for internet outages.

Yes, it’s all doable but I won’t expect it be “set and forget”. I expect there will be quite a bit of tuning with some possible spam and delivery problems while you get kinks worked out.

I posted part of it here: https://forum.ghost.org/t/interested-in-ansible-role-to-install-ghost-to-run-with-podman-containers-systemd-private-networking/46226/2

I also use Ansible. Using Podman’s “quadlet” adapter, the containers run as systemd services.

Congrats on the cat box cleaning!

There’s also Zitadel: https://zitadel.com/

Also, all spam messages.

I host routing for customers across the US, so yes I need it all. There are ways to solve the problem with less memory but the point is that some problems really do require a huge amount of memory because of data scale and performance requirements.

Nope. Some algorithms are fastest when a whole data set is held into memory. You could design it to page data in from disk as needed, but it would be slower.

OpenTripPlanner as an example will hold the entire road network of the US in memory for example for fast driving directions, and it uses the amount of RAM in that ballpark.

Simple means different things to different people.

I self-host Ghost and find it pleasant to use and low maintenance. It is a single Docker container plus MySQL. I recommend a reverse proxy in front of it like Nginx. There are importers from many other blog formats.