Every one of these inspects without modifying anything, and refuses to assert what it cannot point at.
What the system is made of, how requests flow through it, and where the boundaries actually are — as opposed to where a diagram once said they were.
Services, jobs, scheduled tasks, queues, external calls and the things running that nobody remembers starting.
Context and container views built from the code, with every element traceable to the place it came from.
What happens end to end during the flows that matter: sign-up, checkout, the nightly job that everybody is afraid of.
Internal and external dependencies, which ones are load-bearing, and which single failure takes the product down.
Where the cost of change is concentrated, with an estimate of what it is costing you now rather than a generic scolding.
Describe the situation in your own words and see the procedure — and its cost — before you spend anything.