The metal floor shuddered as Zerra crashed
Into what, this is 'incomplete' for a description, into the ceiling, into the wall, the floor. This is likely something you should reorder for 'Zerra crashed into the metal floor causing it to shudder.' Which reads a bit cleaner for semantic completeness. What you wrote isn't wrong, so much as it can be awkward to read especially for people who might not be english first language.
as the spectators clawed the chains.
As another poster asked, what chains, for what you posted there is no establishment of 'chains' before this.
“The fight is rigged!”
“Give us our coin back, you houndspawn!”
“You two-faced cowards!”
In a loud location, especially with chants, there isn't room for fully coherent sentences, thats why short chants and one word yells are common, try thinking of any stadium for sports. I'd recommend staying away from longer multi-sentence shouts, just 'houndspawn' and 'refund' would be strong one word chants.
The chains rattled violently, but the arena fell silent when a few Towering fiends stepped into the crowd.
Again what chains, attached to what, doing what, why are they rattling, are you implying the 'towering fiends' stepping into the crowd did that, is towering actually a part of their name, or a descriptive word you could be capitalising Fiends to make it clear that both are part of the name.
Following this on the next line below:
As their thundering footsteps resounded around, Aegis looked at them.
The enforcers.
personal bodyguards of the owner who runs this hidden arena.
I could be pedantic here, 'whose' footsteps if you mean the towering fiends, why isn't that line with the part before.
It would have made more sense if the footsteps line was leaning on some other line around it, but on a new line it could confuse people, especially since it could be talking about the towering fiends, Aegis or the enforcers, who might be a third party or the same towering fiends. Which is also a note to explain what names and descriptors a person/creature has before using new terms for them if you are switching between a few different characters.
It is generous to even assume by this line that Aegis is even a person, what if they are a faction or group of guards, with how you split up lines currently, Aegis could literally be the title of the enforcers and due to it ending in an s, could already be plural.
Again, this is slightly pedantic, but this is why breaking up text in a more reliable order can help with reading.
Murmurs spread across the crowd as they cursed and spat.
This isn't a directed action with any context, you have a habit of breaking up your lines,
Why are they cursing and spitting, the scene hasn't explained anything about why any character here feels anything, this might be a context issue relying on earlier parts of whatever you have written, but stand-alone this is strange.
Aegis turned around with a sly smile on his face and walked back to his table.
Again a very out of context 'emotion', including how this is the first time this name has properly popped up in this small section giving us even less context, nothing in this scene has been tied meaningfully to this 'Aegis' person so far.
The wooden chair groaned as he sat on it.
Again, a very strange line break, it is disconnected, if you are adding detail or description, keep it tied to something.
“One gold and ten silver,”
To be pedantic again, this seems almost james-bond-esque, it makes little sense so far in the scene and doesn't really build the situation, the person sat down and what, spoke a random amount of non-descript currency to the air because no one else was put in the scene with them?
Who is the listener, what is being paid for, if anything, is there some slave trade happening in the background we haven't been told of, is that Aegis threatening some listener with a price, was it spoken to Zerra to price their freedom or for the steak. There is too much going on for this line to have so little context.
If I had to say something right from the start, I'd recommend trying to follow an old mnemonic I grew up with when I was learning to write:
TiP ToP, [TI]me, [P]erson, [TO]pic, [P]lace.
Thats a basic guideline for when you want to start a new line or paragraph, it helps space things out a lot better and can help with keeping relevant things together.
It isn't perfect, webnovels and general writing let you get away with a lot more than this, but for learning it can help add a bit of structure.
When you change the time of the scene, such as a change of daytime into night, when you change the person you are talking about, when the topic of conversation changes from fish to planes, or when you change locations.
After that, I'd recommend putting some time into figuring out what every detail of your scene is trying to convey, does the chair groan because someone sat in it actually build the scene, are you trying to convey the person is overweight, that the location is poor and can't afford better chairs, or are you adding description to try and fill a gap in the scene.
Given what bits you have put down, we have no clue if Zerra crashed because it is the end of the fight or because they were dropped into the arena to fight, which it didn't read like that initially due to lack of context, descriptions from the establishing part of this scene are vital for readers to understand what the space is, how characters move through it and what it contains.
Without the establishing segment of the scene, thats why so many questions about the chains, about the audience, about the setting come into play.
Us, here in this forum, have no clue where the VIP section is, if 'chains' is meant to be a chain-link fence or something to keep the audience back, if there are any other details about characters, guards patrolling or anything else.