I read every block from launch on 30 Apr to 9 Nov using only the public RPC. No wallet nothing signed

What I found

Finality is real. In 20 out of 20 checks the finalized block and the latest block were the same block with the same hash. So once a block shows up its already final. no waiting for confirmations

Speed is about 1.34s per block not under 1s. Rayls site says under 1s but its own docs table says over 1s so they need one clear number

Since launch the chain had 44 gaps of 60s or more. 24.34h in total so its producing blocks 99.47% of the time. 4 of those were real halts over 30 min and the longest was 11h43m on 5 to 6 May

The weird part is timestamps. I found 203 blocks with old timestamps. On 27 Jul blocks kept coming normally but 32 of them were stamped up to 12h behind. Any explorer that trusts block time would show an 8h outage that never happened. My own first scan fell for it too

Good sign is @Rayls Rayls published a public root cause writeup for the 21 Jun halt and the onchain data lines up with it

My take. Finality is solid and thats what matters most for settlement. What I would fix first is the timestamps. If you build on Rayls order by block number not by time

Full article on my X and all scripts are public: https://github.com/marwan786797/rayls-finality-check

Sources

https://www.rayls.com/

https://docs.rayls.com/docs/axyl-performance-benchmarks

https://explorer.rayls.com/

https://github.com/raylsnetwork/axyl/blob/main/doc/incidents/2026-06-21-mainnet-halt.md

#Rayls $RLS