Disclaimer
I am new to Lightning and started diving deep into it after abandoning it a couple of years back. I learned about swap services and wanted to try them, but I had a problematic experience losing(kinda, not completely) control of my sats(500K) for 30hrs.
Issue
I used a site to swap from LN to onchain, and it worked like a charm without a hiccup. However, when I tried to do on-chain to LN, despite having six confirmations of me transferring my sats to the swap service provider's address, I started to worry. I reached out to middleway(a swap service provider working with Boltz software), and he was invaluable from the start till the end, where he was debugging the issue in real time with me.
Resolution
He had to raise a GitHub [issue] (https://github.com/BoltzExchange/boltz-backend/issues/782) and tweak some code manually(I don't understand the details), and it worked. I finally have my sats back after 30 hours of not having it with me(though I had the recovery private key for the escrow that held the sats, so I never felt scammed or cheated). If that manual tweak had not worked, I would have had to wait 7 days(block height till timeout) to get back my sats. I did not know all this as I was new to Lightning and Swaps, so this was a little stressful, not knowing if I would get my sats back. I still do not understand what technically happened in detail, but I learned a lot from swaps.
Summary
In summary, if a user's feedback for production software leads to a GitHub issue, I conclude that swaps(at least on-chain to LN) are not ready for prime time. The other conclusion is that not enough people are trying these services, which leads to users becoming beta testers. I am too new to LN to give a definitive answer. However, I appreciate the middle-way person sticking with me(having some tough conversations) until the issue is resolved. I have my sats back, so I am relieved. The positives from (non-custodial) Lightning so far have outweighed the negatives, but I feel there will be a few more hiccups along the road, for me at least.
PS: I tried to document the process in real-time in another thread, which you can find on #847217