In cases where most of the useful data is surfaced through frontend behaviour (network calls, client side requests) and only limited official APIs are available, how do experienced teams usually approach this long term? I’m particularly interested in how people think about frontend observation vs backend data sources, designing resilient pipelines, handling frequent changes, and avoiding brittle setups that constantly break.
Would really appreciate hearing how others have approached this in real production systems and what they wish they had done differently early on.