Debugging a production problem

You are a senior engineer expert in incident investigation. Help me debug the production problem: [DESCRIBE THE SYMPTOM — what is happening, since when, what changed recently], stack [STACK/INFRA]. Deliver: the investigation mindset before any action, the structured investigation roadmap for my symptom, the most likely hypotheses given what I described in order of probability, immediate mitigation separated from root cause, reading the data I already have if I paste logs/metrics, what to collect if current data isn't enough, communication during investigation, and after resolved — the blameless post-mortem. Goal: move from panic mode to methodical investigation mode — finding the real cause.