It has been a while, WordPress blog.

I did not come back to this site because I had some grand plan. When I looked again, I found an AdSense email from January 27, 2026. It said the site had issues that needed to be fixed before ads could run. From Google’s point of view, the site had almost no recent writing and was barely being maintained, so the message was fair enough: fix the problems if you want to keep using AdSense.

AdSense email asking for site issues to be fixed
The AdSense notice asking me to fix site issues.

I wonder what would have happened if I had paid attention then.

But the site had been neglected for so long that I let the email pass like any other notification. AdSense was not making meaningful money anyway. Maybe I had lost interest in writing.

A few months later, on June 26, another email arrived. This time it said the AdSense account would be deactivated in 30 days. That was when I finally thought, “Getting AdSense approved was hard, though…”

AdSense email warning that the account will be deactivated after 30 days
The AdSense deactivation notice.

Even after that email, I only reopened rlax.kr about a week later.

On July 2, 2026, I opened rlax.kr again. To be exact, it was less “I should run this properly again” and more “Can I really keep leaving it like this?” The site was technically alive, but almost dead. AdSense would need another review. The old site had become something that was merely costing money.

At first I wondered whether I could patch it up and revive it. That thought disappeared quickly. There were not many posts, and almost no current thread that I could continue from the old writing. Rather than holding onto a few old posts and polishing them, it seemed better to reset the direction of the site.

2026.01.27Site issue noticeAdSense had already warned me to fix the site, and I let it pass.
2026.06.26Deactivation warningThe 30-day notice finally made the old approval feel worth saving.
2026.07.02Opened it againIt was less ambition than the feeling that I could not leave it like this.

The more I use AI, the more I miss my old records

I use AI constantly now, and the more I use it, the more I regret not having more records of my own past.

When I first started using AI, I thought personal data should be handled carefully. I still think that, but these days I feel the importance of my own records more strongly. In the past, writing on a blog felt like extra work, so I did not do it consistently. Now I think differently. If I had left small notes here over the past few years about my attempts, failures, and decisions, they would be much more useful in this AI era.

I recently started using Obsidian too. I had heard many people praise it on YouTube, but I did not expect it to help me that much. Then someone recommended it, and I began using it. Once I installed Obsidian and looked for my own writing or records to put into it, I realized there was not much to add. Like many people, I started with what remained in Gmail and scattered notes, and built something that could almost stand in for me.

Even that was not a huge amount, but it was enough to reveal repeated interests and work patterns. Seeing that possibility made the absence feel sharper. A blog does not have to be filled only with posts designed for traffic or revenue. Work notes, failed attempts, and what I was thinking at the time would have become useful personal data by now.

Obsidian graph view
The Obsidian graph can make records look abundant, but I had very little process and judgment saved on the blog.

What I noticed while creating a new instance

So I decided to start again. It is not too late.

One thing surprised me while creating a new AWS instance: Lightsail had become much easier. Years ago, applying HTTPS took a lot of searching, copying commands into the terminal, and checking certificate settings. I remember saving notes because I thought I would need them later. This time, I could handle much more from the Lightsail menu itself. It may sound small, but when you return to a service after a long time, that kind of improvement feels large.

Lightsail domain and SSL settings screen
The Lightsail domain and SSL settings felt much simpler than I remembered.

Since I was starting over anyway, I also chose a slightly higher instance option than before. The old setup used 512 MB RAM, 1 vCPU, and 20 GB SSD. This time I chose 1 GB RAM, 2 vCPU, and 40 GB SSD. On paper it is not a dramatic upgrade, but the old setup had felt cramped at times. I think I had been trying too hard to save money.

The price difference was smaller than I expected. On the Lightsail pricing table I checked, a Linux instance with public IPv4 was $5 a month for the old configuration and $7 a month for the new one. The difference is $2 a month, or about $24 a year before taxes and card exchange-rate details. Compared with the added comfort, that did not feel like a huge gap.

Lightsail instance settings screen
The old Lightsail instance even had a blueprint support notice. With PHP version issues on top of that, starting fresh felt cleaner than patching the old setup.

Even the last two years were already a loss

I chose the new setup cheerfully enough, but once I opened the numbers, the recent two years alone were already a loss.

If I estimate the old cost from January 2021 to June 2026 at roughly $8 a month, the total comes to about $520. In reality, I remember some months being closer to $12, so that is a conservative estimate.

AWS Cost Explorer made it look worse, not better. It let me view up to 38 months, and even looking only from July 2024 to June 2026, the total was $364.97. The monthly average was $15.21.

AWS Cost Explorer chart from July 2024 to June 2026
From July 2024 to June 2026 alone, the AWS cost was $364.97.

The number left in AdSense, as far as I remember, was about $10. Put those numbers next to each other and it is hard to say the blog earned money. I was simply continuing to pay AWS for server costs. This was not a revenue blog. It was a loss-making blog. Maybe even that phrase sounds too generous.

Recent AWS cost $364.97 from July 2024 to June 2026
Monthly average $15.21
AdSense memory About $10
Real loss Not the money, but the missing record
The cost follows the Cost Explorer period mentioned in the article. Taxes, exchange rates, and billing details may differ.

This time, though, I want to accept that loss differently. Not earning money is obvious. What feels more wasteful is not the money. It is that almost no records accumulated here over the past five years. Even if I had not written many polished posts, it would have mattered to leave traces of what I was thinking, what I did, what I changed, why I stopped, and what annoyed me enough to postpone.

The painful part was not the server bill. It was the empty space where the records should have been.

So I am restarting rlax.kr with a different attitude. It would be nice to pass AdSense review again, and it would be nice if the site eventually paid for itself. But that is not the first purpose. The first purpose is to leave records that future me can open again.

At least this time, I want to record even the fact that I neglected it: the January email that told me to fix the site, the June deactivation notice that finally moved me, the more than $364 spent in the last two years alone, and the strange regret that only arrived once I started using AI enough to wish I had more of my own writing.

That feels honest enough for a first post after reopening.

Keep Reading

If you want to know what kind of archive this site is trying to become, start with About and Editorial Policy. From here, I want to keep recording observations, judgments, things I used, and things I missed one by one.