as quoted before, have opened port 80 and http services.
Yes, I read that, but simply opening the firewall rules might not have been enough. Thus my request that you
test it from the outside by going to some network on the other side of the Internet and try to HTTP back in. That would've been enlightening had you done it while the symptom was occurring. Now it's gone, so you're left with a mystery.
Let's Encrypt was ignoring my request.
Which is more likely: your idiosyncratic one-off local setup had a problem or a world-scale infrastructure service used by and depended upon by millions had a blip just at the moment you needed it to work?
Huge services do sometimes go offline, but usually when we think of such things, it's ones like Google, Facebook, Amazon, and such, which we expect to provide ongoing service for hours at a time. Let's Encrypt does its work in moments, and no more than once every few months. Purely on statistics, therefore, we should expect that Let's Encrypt "always" works. Close enough as to not be worth blaming without absolutely damning evidence, anyway.
I don't exclude myself from this logic. The last time I had a failure related to Let's Encrypt, it was a combination of an OS vendor's mistake and a weakness in my own local configuration. Let's Encrypt itself was working just fine.
Randomly, it's working. Nothing has changed.
This is a sign of trauma, alas. You've been taught that computers are nondeterministic and can "just" start failing and then start working randomly.
RouterOS isn't perfect, but it's more reliable than consumer OSes. Expect better from it. Chances are much better that any given problem is either some hard limitation you can fix or work around, reliably and durably, or it is a local mistake in use of the tools.
In this particular instance, I suspect the application of this principle is that the problem never was Let's Encrypt or RouterOS, and that the intermittent problem wasn't properly fixed and so will come back again.
My vote? Your dynamic DNS to IP mapping was out of date and got updated automatically in the background. No accurate DNS = no Let's Encrypt. If that's the case, it means your symptom will recur the next time your router updates the cert while the dynamic DNS is out of date. Alas, with LE updates being months apart, that's likely to happen after you've forgotten this conversation, so it'll surprise you again and you'll be wondering why HTTPS "suddenly" went down "for no good reason."