These days, on-chain data keeps "hanging up," and my first reaction in the group is to curse the slow chain... Actually, many times it's not the chain itself, but the indexer/subgraph tracking blocks, encountering rollbacks or reorganizations that require recalculating; plus RPC rate limiting, during peak times your requests queue up, and the charts seem to be paused. When the market-making dashboard stalls, it's easy to get anxious, and now I’d rather wait two minutes than chase orders blindly. During mainnet upgrades/maintenance, it’s even more obvious, and everyone starts guessing whether the ecosystem will migrate, to be honest, first check if the data sources are the same set. Anyway, I no longer believe that "what I see is real-time."