Ask HN: Solo founders, what’s “stuff breaks while you're unavailable” strategy?

#1

Post by leros » Mon, May 23, 2022, 12:40 AM UTC

There is a current post with a similar title asking about your "get hit by a bus strategy" and I think the answers are missing the mark, so I'm posing a slightly different question.

There are times you are absolutely unavailable (sick, on a plane, off grid, dealing with a personal emergency, etc). What's your strategy when all hell breaks loose and you just can't step away.

In other words, being on call 24-7-365 is not realistic. Do you have a plan?

Re: Ask HN: Solo founders, what’s “stuff breaks while you're unavailable” strategy?

#2

Post by efortis » Mon, May 23, 2022, 12:51 AM UTC

For example, a Server's:

- disk fails (ZFS with hot spares)

- power supply fails (each server has two)

- someone cut the power (there's another server)

---

Most issues will be caused by me, so the general answer is don't leave before testing.

Re: Ask HN: Solo founders, what’s “stuff breaks while you're unavailable” strategy?

#3

Post by devKnight » Mon, May 23, 2022, 1:12 AM UTC

efortis wrote: post #2

For example, a Server's: - disk fails (ZFS with hot spares) - power supply fails (each server has two) - someone cut the power (there's another server) --- Most issues will be caused by me, so the general answer is don't leave before testing.

Not a solo founder, but this is a huge part of it. Do not deploy changes to code or architecture unless you CAN be on call to fix what breaks.

Re: Ask HN: Solo founders, what’s “stuff breaks while you're unavailable” strategy?

#4

Post by leros » Mon, May 23, 2022, 1:18 AM UTC

efortis wrote: post #2

For example, a Server's: - disk fails (ZFS with hot spares) - power supply fails (each server has two) - someone cut the power (there's another server) --- Most issues will be caused by me, so the general answer is don't leave before testing.

Do you not worry about something breaking even with an older deploy? Random things happen.

Re: Ask HN: Solo founders, what’s “stuff breaks while you're unavailable” strategy?

#5

Post by efortis » Mon, May 23, 2022, 1:30 AM UTC

leros wrote: post #4

efortis wrote: post #2

For example, a Server's: - disk fails (ZFS with hot spares) - power supply fails (each server has two) - someone cut the power (there's another server) --- Most issues will be caused by me, so the general answer is don't leave before testing.

Do you not worry about something breaking even with an older deploy? Random things happen.

I don't worry about random things.

Re: Ask HN: Solo founders, what’s “stuff breaks while you're unavailable” strategy?

#6

Post by ilaksh » Mon, May 23, 2022, 2:19 AM UTC

I am really looking forward to the point where it doesn't take all day for someone to show up in my Discord to mention the website is down.

Until then it's apparently not a huge problem.

Re: Ask HN: Solo founders, what’s “stuff breaks while you're unavailable” strategy?

#7

Post by Matthias247 » Mon, May 23, 2022, 4:47 AM UTC

leros wrote: post #4

efortis wrote: post #2

For example, a Server's: - disk fails (ZFS with hot spares) - power supply fails (each server has two) - someone cut the power (there's another server) --- Most issues will be caused by me, so the general answer is don't leave before testing.

Do you not worry about something breaking even with an older deploy? Random things happen.

Yes, e.g. hardware of hosts breaking, datacenters getting disconnected, etc.

I'm not a solo founder, but if I would be I would most likely pick a solution where I would have to maintain the least amount of infrastructure (e.g. using SAAS/FAAS offerings). That would mean in most cases someone else gets paged if things break. Still doesn't guarantee 100% availability since it can't prevent code bugs from doing bad things, but there would be still less things to worry about.

Re: Ask HN: Solo founders, what’s “stuff breaks while you're unavailable” strategy?

#8

Post by hutrdvnj » Mon, May 23, 2022, 9:06 AM UTC

Risk acceptance. What happens if you die tomorrow by a car accident? Accept that risk and switch focus to more important things.

Re: Ask HN: Solo founders, what’s “stuff breaks while you're unavailable” strategy?

#9

Post by leros » Mon, May 23, 2022, 1:21 PM UTC

Matthias247 wrote: post #7

leros wrote: post #4

Earlier quoted context omitted.

Do you not worry about something breaking even with an older deploy? Random things happen.

Yes, e.g. hardware of hosts breaking, datacenters getting disconnected, etc. I'm not a solo founder, but if I would be I would most likely pick a solution where I would have to maintain the least amount of infrastructure (e.g. using SAAS/FAAS offerings). That would mean in most cases someone else gets paged if things break. Still doesn't guarantee 100% availability since it can't prevent code bugs from doing bad thin…

This is my take. I build products that have the critical infrastructure be fairly simple and hosted completely on third party systems. The most custom part of my system is just the customer portal, which can go down without impacting the core product functionality.

Re: Ask HN: Solo founders, what’s “stuff breaks while you're unavailable” strategy?

#10

Post by patiiipot_ » Wed, May 25, 2022, 12:59 PM UTC

Matthias247 wrote: post #7

leros wrote: post #4

Earlier quoted context omitted.

Do you not worry about something breaking even with an older deploy? Random things happen.

Yes, e.g. hardware of hosts breaking, datacenters getting disconnected, etc. I'm not a solo founder, but if I would be I would most likely pick a solution where I would have to maintain the least amount of infrastructure (e.g. using SAAS/FAAS offerings). That would mean in most cases someone else gets paged if things break. Still doesn't guarantee 100% availability since it can't prevent code bugs from doing bad thin…

My team from Founders Cafe told me to create a strategy, kind of like a flowchart of what to do in case I'm not on call when there is a problem. I give them to my employees/freelancers. As soon as I am available, I'll check the conditions of the workplace first thing in the morning.