Recently, many people have been saying that on-chain data is "getting stuck," but in many cases, it's not that the chain has stopped; it's that the data pathway you're looking at is congested: the front end queries the subgraph, which is backed by an indexer tracking blocks and writing to the database; at the same time, wallets/webpages also need to make RPC calls. If one side is rate-limited, or if there's a reorganization rollback, or if node responses are slow, the page just spins like crazy, then after a few seconds, it’s fine again. Basically, the on-chain world also has "caching + queuing."

Modularization and DA (Decentralized Applications) narratives break down the architecture into finer parts, which developers of course love, but regular users get even more confused: why is there another layer, and why do we need more services to be stable... Anyway, when I see data latency now, I just assume "it's not an opportunity, it's a risk," so I reduce leverage where I can, and focus on survival. Today, I got annoyed again by an RPC rate-limiting popup, really.