I've lately decided to set up a 1Password Family Account to help at least my nuclear family into taking net security more serious.
- What steps did you take to make it simple enough for your family to care?
- Did you retain any restorative powers? As in keeping master passwords to certain things and/or emergency accesses like in LastPass?
- Which subjects spurred the most discussions and how did you solve it?
- Which items do you share amongst all family members?
Edit: Formatting
If it's not a government agency, it doesn't get any real info.
So, for instance, Amazon has a made-up name and our PO BOX and our "junk" number from Twilio.
As I have mentioned here many, many times:
This is possible because VISA/MC do not verify cardholder name. They make it seem like they do and merchants think that they do ... but they do not. You can just enter "Mickey Mouse" and it will work just fine.[1][2]
Lyft, opentable, Toasttab, Apple ... none of them have ever seen our real names or addresses.
[1] This is not true of AMEX - they do verify cardholder name the way people think they do.
[2] There is a very rare, seldom used "verified by visa" step that some online merchants used to use (mostly in Europe) that did verify cardholder name ... but I have not seen it in years ...
If we assume this information persists forever - and we might as well - it represents infinite liability and risk whereas the mitigations I have proposed cost almost nothing.
Or, to put it another way, it's very cheap insurance.
You can call it paranoia. I'd call it a healthy value of privacy.
Nothing stops a company from taking your name from your credit card and using it to build/sell shadow profiles except their word. Companies' words aren't worth shit.
The data is there, silently, stored for a future use which may come at an unpredictable time.
If you're worried about the government prying, using a fake name on a PO box is a great way to rouse suspicion and investigations into whether or not you're doing something nefarious with the post.
The post office (quasi-government agency) knows who we are.
I don't think rsync is worried about government prying.
We live on a ranch at the end of a 2 mile private road.
Lyft is something we use in the city or while traveling.
However, I do understand the spirit of your question and I don't have formal practices for these kind of things.
Like a sibling comment said, life is short ...
Beyond that, for the type of tracking and fingerprinting done today, how relevant is your name actually?
Doesn't even have to be that big of a policy change; if Facebook gets suspicious about you using a fake name or various other scenarios, they'll lock your account until you provide a scan of a passport or state ID.
Sad.
For most things though I dont mind too much about things that are already public anyhow. Like name, address etc.
That said for services that extend beyond that, If i feel the need to go that far i just dont participate. Like facebook for example.
I do carry a google voice number. That is the only thing a company, service or basically anything that gets typed into a computer gets. It has signifigantly cut down on spam calls to my actual cell phone, which friends and family do have.
The only really private way to use these services would be using something like burnermail.io and crypto.
They get firefox browser with adblocker preinstalled. I manage their important passwords (eg fastmail) and trained them to rely on firefox sync for the recoverable accounts.
I use MeshCentral for remote administration (amt)… its amazing for the price (free).
OPNsense firewall in all the homes. Unknown devices are isolated and egress over wireguard VPN.
Only apps that truly need admin rights (that install services, etc.) would be blocked. Everything else is wide open.
The admin rights restriction on app installs was almost just a convention that people followed. Now that the app incentives have changed (malicious apps don’t try to take over the machine anymore, they just try to steal your data), admin rights restrictions are becoming irrelevant.
> No microsoft accounts.
I'm asking because I do exactly the same: no admin accounts for my family on their own Windows machine (mother in law and wife etc.).
They're also using only local account. Is Microsoft banning local accounts?
We will likely release some of these restrictions as our trust in our child and their recognition of the problem sinks in over the next few weeks. I would never consider an outright ban or prohibition because I agree it does have some of the affects you mention. Instead, this is a resetting of expectations for our child around what level of risk is acceptable.
I remember being in IRC chatrooms and having predators galore trying to get me to hang out with them. From my understanding, the problem hasn't gotten any better.
A parent would be smart to limit some internet access to their children. We ID at bars and brothels (where legal), the way I see it, the day my son is willing to learn to jump through hoops to access pornography is the day he is ready to watch pornography. I'm not all too concerned about them working around these things, but I'm concerned about easy unrestricted access.
I did the same thing as in the office - embarrassed or annoyed them (in a small way) by using their lack of security. I changed desktop backgrounds, "stole" £20, sent emails with promises like "I'll wash your car" to people. I'd follow this up with a lecture on "if I can do this, imagine what some dodgy foreign hacker could do".
Constructively, I pay for the whole family's 1Password and Fastmail accounts. I am the admin. I'm patient and understanding when they do something wrong. And I limit the number of people I help to those I can really help.
We have a WhatsApp group where they can ask whether something is dodgy. They don't use it for chitchat, so anything that comes through, I treat urgently.
I really like that idea.
I'm a big believer in Signal (I have a monthly contribution set up) but Whatsapp is far more common for most people to use.
I have convinced my older parents and siblings to all use Signal though, so I consider that a win.
+ Network is covered by pihole (and in exchange, plex/jellyfin/etc access works nicely)
+ Smart home stuff is managed by me. Everyone has admin rights but shared terminals (eg kitchen panel) are unpriviledged users.
+ Everyone has a home directory on the homelab they can back up to with as much space as they want (4tb+). I help them set it up if they ask.
+ Haven't done this yet but would like some kind of network level monitoring for threats (viruses, cryptominers, etc)
Things intentionally not done:
+ I don't install anything on folks devices.. at all, but never without their consent and without them having an off switch.
+ We have cams but everyone can turn them off and view recordings. Recordings are kept only for a short timeframe. Cams are all visible/known.
+ I intentionally collect no logs of dns or other stuff. When I do occasionally need to debug an issue, I let everyone know I am flipping on logs for a few minutes.
Empower users, don't control them.
I install whatever I can to control/centralize the devices of my wife and parents. The less they know the better. Because your know, I am the 24/7/365 all-knowing support.
I do not do this with my teen kids. They can manage their stupid themselves. Until recently I had them tracked in Google maps but not anymore. They do see me though.
I have found the opposite to be true. If I push them to invest and understand, they are more likely to fix their own problems. I play tech support very infrequently and usually it's just initial onboarding - "Hey, how do I watch movies?"
My users can reset their own passwords, reboot devices, and some of them can even restart stuff on the server.
There's a dashboard with all of our links, so I don't get the "What's the url for..." stuff.
I keep quick docs in the family notes.
Usually my only problem makers are game servers since those are always a bit less than stable once loaded up with mods.
I do not go around imposing my beliefs upon them. They have their own problems to be bothered with mine too.
Household shared threat models include:
- they download a trojan, it infects your shared printer, and copies all your documents to the bad guys
- their phone is compromised and its microphone listens to your conversations and keystrokes.
These are attacks against high-value targets. If you're just surfing Reddit, you probably don't care. But if you work for a tech company and have customer data or cloud logins on your machine, you should probably care.
Printer could be an issue, I guess, but most tech jobs don't involve a lot of printing and if they do a USB only printer should be supplied by the company.
- We have a phone that never leaves home and has no SIM card. We use it for banking apps and 2FA critical services.
- So our "street phones" don't have any banking apps installed, nor social media apps, 2FA nor password managers.
- We have a paper notebook with secrets and 2FA recovery codes in the bookshelf sitting among many other notebooks and old dusty random stuff.
- Our kid's phones have DNS pointing to Cloudflare's family filter server. Their YouTube accounts are set to filter adult content.
- We use BitWarden family plan for sharing passwords among us.
- We use a Keybase team to share documents between us and our personal devices. Everything is, in theory, encrypted and we can revoke the device in case it gets lost/stolen.
- Our laptops have luks drive encryption and we transport them turned-off. So in case they are lost/stolen, data in the drives are unreadable.
"Smishing" (aka "SMS phishing"), for a start.
Then it seems to me that a phone that is only used for banking apps and 2FA is less likely to be owned than a phone used for everything under the sun.
Email i use on a gsuite legacy domain and have for a very long time. It allows us to move email around if needed. We still have some older gmail accounts as backup, but rarely use them. Ill probably move to something else, Mail in a box on a linode or protonmail. The problem is i havent found a 1:1 feature, between google voice for voicemail and junk phone #, and contact syncing.
On the network i manage that. Use opnsense with unifi for wifi and a few vlans. We dont have cable, so roku's/ROKU tvs get their own DMZ and we have plex and a few streaming services.
I also help manage my parents network. So they have a pfsense appliance (setup and bought well before all the nonsense) and it has a VPN connection to my house, with a similar Unifi wifi network.
All of our stuff is MFA enabled and i just handled the setup on her phone etc, gave and setup yubikeys etc.
Outside of my parents and wife/family, i dont really get involved. I really dont want to. My in-laws I have helped do things for like setup some wifi extenders etc. But their needs are more simple and dont require the complexity my parents do (that WFH and run a business from home with a larger layout.)
I will say one gotcha that got me...The dont allow Google Voice for workspace. If youhave a legacy account, even on the gsuite it should be fine, but you cant setup new service. And if you mark the payment account as "individual" you have NO options. If you set it as a business you could get google voice as a paid service.
Payment accounts cant be changed once setup. Which is crazy.
Mine was almost 20 years old.
I have already moved to workspace for now. There will be no charges until August or so. And then 50% off through August of 2023. So its....not awesome but gives me breathing room to find an alternative (Or just stay the course).
I will say one gotcha that got me...The dont allow Google Voice for workspace. If youhave a legacy account, even on the gsuite it should be fine, but you cant setup new service. And if you mark the payment account as "individual" you have NO options. If you set it as a business you could get google voice as a paid service.
Payment accounts cant be changed once setup. Which is crazy.
In this way, people in my family get to choose their own pathway online. They're informed, and they get to make informed choices about what data they care about protecting vs sharing. You can do whatever you want on your own devices, but not on my network, that's mine and I get to set the rules there. For the most part, folks choose my network for protection and performance rather than the wider freedoms of mobile data. I also provide everyone with a VPN account for their mobile devices for when they're out of the house, and most of us use it, but it's entirely optional.
"Before crossing the street, look both ways. Or don't. But you might have to get a job to pay off your hospital debts. I'm respecting your autonomy."
I don't feel it's appropriate to write more thoroughly on the topic, but it suffices to say that respecting autonomy and privacy and making age-appropriate loosening of restrictions is, in my opinion, important for the development of personal responsibility, critical thinking, and life skills that serve in adulthood and crucial to childhood development. Parents are responsible for safety and guidance, and as children get older, the slider moves more towards guidance and less towards safety. I allow my children to do all sorts of "risky" things if they do so in an informed manner.
All secure tasks like handling of IDs, banking, trading, etc. must be done through managed Linux workstations (Landscape with master image), or managed VDI. Keepass is used to store credentials. There is a network storage accessible only to those workstations containing important documents. A second storage area is avaliable for unmanaged and Windows devices.
Windows devices have Group Policy set for update settings, but generally users can do whatever they like. Mobile devices are expected to be patched but they have free reign. Haven't found a good management solution for Windows and mobile yet.
Wifi uses EAP-TLS, no exceptions and no guest devices permitted. As a result IOT and smart home devices are not allowed on the network since they don't support EAP-TLS. Certificates are issued per device and allow access to different services like VPN etc.
I currently don't have managed switches so mobile devices and personal workstations do share the same network as my servers and such, but all local services like network storage are encrypted and require authentication. Ideally I'd have VLAN segregation, but this will have to wait for the next network upgrade.
I back up (export) 1Password vaults quarterly to an offline backup I maintain.
I maintain two small (1Tb) SSDs with digital copies/scans of all important documents, offline. Try to sync monthly. Store inside faraday bags inside fireproof (in theory) safes.
We both lie excessively when creating profiles for online accounts. Unless absolutely necessary we use a PO Box for addresses. We've both been online 30+ years and the amount of old, forgotten accounts that resurface in breaches and scams is disappointing, and yet not surprising. Our late dog continues to receive a lot of "growth hacking" spam from services that started after she died.
Everything important has multiple 2FA options enabled, avoiding SMS as much as possible.
All of our financial accounts use email addresses off an obscure domain name I manage, not our personal email addresses (which themselves are G Suite/WorkPlace/WorkSpace/Whatever accounts).
I review all financial accounts monthly to look for odd charges. The last serious fraud we experienced started as small (<$10) charges over several weeks, I guess testing the credit card information they'd gotten.
I just assume we'll get hacked at some point, instead of trying to make that impossible, I try to ensure that we have backups of everything (and a paper trail as necessary to prove who we are, though I'm not convinced the various automata at FAANG gang companies will believe any of that).
For our purposes it’s been fine, but it’s overkill for the typical family or typical consumer.
One definite downside is that G Suite accounts are not considered to be consumer accounts so you run into various Google services which either don’t work at all or work very differently. For the brief time we used Google Home it could not access either of our G Suite calendars (but somehow the Alexa could). Our Nest footprint exists in a separate world from our G Suite accounts. When we had YoutubeTV we had to use a separate GMail account because (at the time, I don’t know if this is still the case) …because G Suite accounts could not be used for YouTube TV.
I had convinced my wife (who is not of the tech world) to switch to a password manager a couple years ago, and while she didn't love it, she's now totally on board.
My mother runs a fairly successful small online business and kept getting BS charges on her business cards along with other various occurrences. She and her employees were sharing a couple passwords for everything the company used. They weren't _bad_ passwords, but it wasn't a great set-up.
Finally, my wife and I convinced her to try out a password manager. After quite a few excuses why it would never work for her over a couple weeks, she got a another bogus charge and begrudgingly accepted. My wife went and spent a day to help her move _everything_ over, generate strong passwords, and showed her how to set up accounts for her team and share them so her team could do the same.
I'd since forgotten about all of that, as it's been well over a year since all that happened, and I assumed she went back to her old ways. Then last weekend I heard her bragging to a colleague about using a PW manager and how it's changed her life significantly - recommending they do the same.
She said excitedly "I don't even know my passwords! It's great!"
It isn't worth the drama frankly and they are grown adults.
>Which subjects spurred the most discussions and how did you solve it?
We had one distant family member go off the deep end with 5G vax/google/phone is listening to me. That triggered a family wide discussion on this. Tried injecting some facts, but its quite hard explaining concepts like fingerprinting and data brokers. Its such a nebulous concept and if you're not careful you just end up affirming the paranoia unintentionally. "oh so you're saying they do track me and read my mails?" Yes, but also noooo
As an example to this, a lot of 9/11 were founded on the "they fell too fast" aspect and "jet fuel doesn't melt steal beams." Both these things are true! The problem is not understanding some basic engineering principles. In this case 1) skyscrapers are have a designed failure mode to collapse in on themselves (and quickly) as to not destroy surrounding buildings if they fail. 2) Jet fuel may not melt steal beams, but it is hot enough that it can cause significant structural weakness, enough that the weight of 1/3rd of a skyscraper will cause said beams to break. Understanding this makes it a far more likely scenario than Bush allies placing hundreds of thermite bombs all around the WTC without anyone noticing.
But the problem is that there was something odd that doesn't fit general knowledge. Knowledge is 100% the cure to conspiracies.
Worse, I think hiding information will just build on the paranoia. Especially with something like data privacy, since we all here work with data. It will turn into "you hid this from me, so you're part of the conspiracy." In that case, it removes all chances that you have. Yes, discussing the nuances of the "conspiracy" can cause people to go a little further off the deepend, but it can also form a bridge between you and that person (as opposed to isolation, which is what most conspiracy theorists do. Put themselves in a bubble). That bridge can be used to lead them out, because they now trust you and you have expertise. It's not easy, but it's something I've personally done.
There seems like a good OSS project - a bit of WMI a bit of bash - so I am interested if anyone has a idea.
You got bigger problems after Password Managers.
Just polished up the transparent Squid/SquidProxy/custom-ICAP-servers-to-block-DNS-over-HTTPS/Default-Deny-firewall for my home.
It seems to me that we are losing the war on Zero-Trust home-based content filtering (with the onslaught of Webroot port 7777, and DNS-over-HTTPS, and even AVG 443 for DNS.
You all hear me? I am (and probably we are) losing control of the HomeLAN/home-net via the onslaughts via circumventions of Zero Trust Model.
And this new DNS RTYPE SVCB and HTTPS by Akamai CDN, Apple iPhone/iPad, Cloudflare, and BigIP/F5 is making this gateway (and me) losing it all.
https://docs.diladele.com/tutorials/transparently_filtering_...
https://datatracker.ietf.org/doc/draft-ietf-dnsop-svcb-https...
Jokes aside,
> making sure they have control of their passwords and accounts in a safe matter.
This looks like two requirements. Control your passwords and accounts and safeguard them. Because saving via Chrome, though unpopular, is quite safe but you give up control.
I've found this to be useful lately as I go through and take control of my login credentials:
https://mullvad.net/en/blog/2021/11/15/forget-your-passwords...
Been operational for a few years. Minimal maintenance. Great peace of mind.
If anyone asks me anything about security I tell them to assume everything everywhere is spying in every way(I don't advise against using any specific devices or services, or avoid them myself, whether they care about Google spying is up to them).
But mainly I just tell everyone that they should be using 2FA. Everyone even remotely tech savvy these days knows that spying us the business model for half the internet, and only a few care enough to do anything about it.
The only in depth discussions I've had(Outside of work of course), have nothing to do with insecure systems and everything to do with public posts.
For everyone who gets their bank details hacked(And probably gets most of it back), there's probably 5 who lost jobs or friends or opportunities, or just embarrassed themselves, because they posted something on a medium that is easy to misinterpret and encourages posting without thinking.
In the last year I know one person who was hacked. They didn't have 2FA on.
These days, if someone's needs can be met by ChromeOS, then they can be met with desktop Linux and a browser, too. Compared to Windows, the support issues almost cease upon upgrading to Linux, as it is virtually impossible for someone who doesn't know what they're doing to break a Linux install. The slim amount of issues I've encountered can be fixed with a restart.
Everyone else: not my circus, not my monkey.
I have done zero or little past basic configuration and have had no issues or surprises.
I also put a basic 2-bay Synology NAS in the basement, and everyone's laptops are set to back up via Time Machine automatically.
* iPhone SE
* Gmail account
* WhatsApp account
* Everything set to auto-update
* Good passwords, written down on paper kept safely
* Never install any apps without me
* Call me if you are ever worried about any email or message
Also, got some yubi-keys which I use for aws and gmail. Still have a raincheck for my wife to try those.
Yes, I told her it's not hard to get started, she could do it herself if she wanted to, but she wants to discuss it at length and spend time on it (I agree with that, but geez, it's hard to do stuff like this). Taxes are a nightmare enough for us.
I knew when I saw the title here that there would be multiple stories of folks who have gone ABOVE AND BEYOND AND THEN SOME, HN-style. Ain't going to happen in my house.
Steps to make it "simple" - use password manager - store shared and individual pw DBs on a NAS where family has access - use Syncthing to keep changes aligned between devices - configure all browsers and devices to be integrated with pw manager and demo proper usage - everything important stored on a NAS that is in my physical possesion and which uses redundant storage (RAID) - implement backup of critical NAS data - test backups monthly! (can be restored? are still occurring properly?) - install a Linux distro and configure key-based auth (my key trusted) SSH for family members who are willing to use Linux on the desktop. 2022 is the year for it! ;) - ensure things auto-update - if problem occurs shell access is a few keystrokes away - can manage family's digital situation remotely to some degree this way. very helpful! much better than the ole' "Call up grammy and try to drive her clicks and typing remotely..." routine!
Restorative powers retained? - yes, except for the master password to any private password DBs
Which subjects spurred the most discussions and how did you solve it? - Linux: Some people have no idea what an operating system even is, let alone how a "Linux" differs from an "Apple" (not OSX, it's an Apple!) or a "Windows". This was solved by reminding them what I do professionally and them remembering how much time I spent behind the screen doing the bits n bytes. Basically "I got u fam, don't worry about it." was my solution. - Social Media: This is an unsolved problem. Some insist on having FB, Insta, whatever installed on their cellphone! It's nuts. I'm not cool with it, but we all make our own choices. I try to educate people on this topic, but it's an uphill battle.
Items shared for all family members - none - within household: shared pw database with things like streaming & delivery service logins, etc
opening a malicious pdf on their main machine or a malicious website
the one time all their sensitive info compromises their main gmail/apple account
How is it even possible to help our extremely vulnerable elderly parents and then our very young family members, nephews, nieces
We've probably all been pwned at least once, and we're the more cautious/aware of the population, how do the helpless even fare? Besides locking them down in the apple eco and idk vetting every file/website they use?
- run my own DNS and tunnel into the home network,
- no TVs,
- no smart devices,
- networked devices in communal spaces only.
I think all the rest like password managers and such are personal choices, but those sorts of behaviors will be encouraged.
There's a line between trying to control the behavior of your family and keeping the environment they're in healthy and safe. I wouldn't want to have a master password or access to all their personal accounts.
My kids get locked down OS's and games, in addition to communications limits and screen time restrictions. But they're elementary age, so this is okay. The rules relax bit by bit as they get older.
For my extended family? Nothin. They're grown ups. I do host the family e-mail domain but there aren't any rules around that (well, they do have to pay for it...). We've had discussions about best practices, but the non-technical folks don't care ("so what if Google tracks me, I don't care") and the other half are technical and more than capable of managing their own digital lives.
You can be non-tech savvy, use Chrome, Gmail, Drive, etc... and get good cloud services that are secure.
Internet safety, DNS security, https://wrdrd.github.io/docs/consulting/kids #internet-safety #family-media-plan #screen-time-guidelines
Rclone supports encryption over top of like every cloud storage provider; and then what js could hit delete and confirm on our cloud storage, resulting in starting over from zero, like preppers, like bushcrafters - with DR bushcraft knives with flints (and hand-crank solar rechargeable FM/WX radio USB powerpacks) - like a low-budget made for TV Swiss Family Robinson: https://wrdrd.github.io/docs/tools#rclone
Ansible-molecule, DevSec baselines; your (1) Raspberry Pi SD card will fail, and probably before a thumbdrive or an SSD.
E2E: Cyph, Keybase has encrypted git repos; GitLab/Gitea does Issues with trackbacks: https://www.cyph.com/blog/cyph-pgp
PWD generates a printable substitution box: https://github.com/westurner/pwd
SGP: SuperGenPass https://github.com/chriszarate/supergenpass
JS implementations of SSS to do better than splitting a string in parts and printing some redundantly: https://github.com/topics/shamir-secret-sharing
"SLIP-0039: Shamir's Secret-Sharing for Mnemonic Codes" https://github.com/satoshilabs/slips/blob/master/slip-0039.m... :
> Shamir's secret-sharing provides a better mechanism for backing up secrets by distributing custodianship among a number of trusted parties in a manner that can prevent loss even if one or a few of those parties become compromised.
> However, the lack of SSS standardization to date presents a risk of being unable to perform secret recovery in the future should the tooling change. Therefore, we propose standardizing SSS so that SLIP-0039 compatible implementations will be interoperable.
> What steps did you take to make it simple enough for your family to care?
GNU/Linux desktops for all, for me NixOS/Emacs (EXWM), for relatives mostly Gnome SHell (the second capital is NOT a mistake, but they want something like that) and XFce, no wifi, at least I have few MikroTik APs but powerd off, powered on only if I have a guest and he/she can't use wired ethernet. Desktops have "proper" WebVM [1] with user.js/various extensions etc all regularly kept up to date backed up and casually restored around once or twice a year when I upgrade from a major release to another. IoT stuff (domestic p.v. + related tools) offline on a separate network with a homeserver (Home Assistant pip-installed, not the absurd docker image) bridging the WebUI part from the desktop's LAN.
> Did you retain any restorative powers? As in keeping master passwords to certain things and/or emergency accesses like in LastPass?
I have a printed copy, "encrypted" with a simple letter substitution scheme those who need know it, of some passwords, so they can ask for help someone who know GNU/Linux if I have some health issues/I can't really help for some reasons, but it's not much a tested setup just something do and explained a bit without really having ever used it so I can't really know how much it can work, it's a potentially serious issue but so far no one seems interesting in that, I'm healthy etc so...
In iron terms I have enough iron to survive various faults on both desktops and homeserver/mini-small-rack side, in software terms everything is almost reproducible with org-mode documented and tangle-ed NixOS configs and relevant custom ISOs ventoy-deployed locally or deployed via LAN depending on the case. Not everything is fully covered but it's enough.
> Which subjects spurred the most discussions and how did you solve it?
Well... The "family policy" a bit against my will is "you are the techie, we do not care" so there aren't really be discussions, just few explanations/training etc
> Which items do you share amongst all family members?
Phone system (Grandstream UCM PBX + GXP phones simply because when my old Asterisk card die and I see an offer for the PBX I was a bit tired of Asterisk), video surveillance, witch is only outside and physically powered off when someone of us is at home. Aside the small p.v. system witch, sigh, is to be counted in the "digital" things since it's full of FWs and to be effective enough (like piloting the hot water production depending on the sunlight) it demand a home assistant...
Essentially my general policy is:
- restricting as much as possible the attack surface
- restricting connected stuff (witch count in the attack surface) as much as possible, still leaving a bit of comfort
- be reproducible
- have a bit of redundant gears, not for anything, too expensive and demand too much space, but for something yes. For instance a VoIP spare phone + two analogs (with the PBX that have two fxs ports), around a desktop (ssds, mobos, CPUs, ram etc) and a half as spare parts, two 16 ports spare switches against a 48 ports in production one (not all ports used, of course) etc.
[1] monsters mostly called browsers for legacy reasons, like Firefox or Chromium that actually are not much more "browsers" than a JDK...