You might be interested in reading Rethinking Lightning if you're really interested in the details:
Channel liquidity is a problem, but it is also deceptive. When you have 100k sats of inbound liquidity you would think you could receive up to 100k sats, but this isn't the case, often you can't actually receive any. This is because of on-chain fees, when a payment is being made in lightning you are creating pre-signed transactions that have outputs for every in-flight payment, these outputs cost potential on-chain fees and the high on-chain fees go the more it eats into your liquidity. After we've solved most of our force close issues Mutiny this has been number one support request. Even if you do everything right, understand liquidity and have enough for your payment, sometimes it still won't work because on-chain fees are too high. This is always really discouraging because isn't the whole point of lightning to not have to pay on-chain fees? Fundamentally, all current lightning channels could become entirely useless if on-chain fees went high enough because a single payment would require too many reserves. Obviously this is hyperbolic, but I hope I am getting the point across that on-chain fees don't just effect the opening and closing costs of channels, even if you are a diligent node runner that only opens channels when fees are low, that is not enough, your channels need to be large enough to pay for the on-chain fees of each HTLC at any future on-chain fee rate. As on-chain fees go up and up this problem will only get worse.