Providing context requires someone to understand the context and the variables that went into the decision (including potential software bugs or edge-cases) and confirm the decision to the standard they're claiming "we've carefully reviewed this decision/etc BS".
Problem is that in reality, the decision is made by some heuristics that can be flawed due to software issues and/or eventual consistency (and the developer who brought up these issues got told "don't worry, it'll only affect 0.1% of users"), and the "escalation path" is an outsourced idiot that presses random buttons while moonlighting as a tech support scammer.
Problem is that in reality, the decision is made by some heuristics that can be flawed due to software issues and/or eventual consistency (and the developer who brought up these issues got told "don't worry, it'll only affect 0.1% of users"), and the "escalation path" is an outsourced idiot that presses random buttons while moonlighting as a tech support scammer.