[NO-TICKET] Fix FallbackAny steep check

[vpellan]

What does this PR do?

RemoveFallbackAny exclusion from the Steepfile and fixes related issues.

Motivation:

Improve typing on dd-trace-rb.

Change log entry

None.

Additional Notes:

How to test the change?

bundle exec rake steep:check

[pr-commenter]

Benchmarks

Benchmark execution time: 2026-01-09 14:00:31

Comparing candidate commit 8a2cb5d in PR branch vpellan/steep-fix-fallback-any with baseline commit 5ec9115 in branch master.

Found 3 performance improvements and 1 performance regressions! Performance is the same for 40 metrics, 2 unstable metrics.

scenario:line instrumentation - targeted

• 🟩 throughput [+10723.525op/s; +11314.036op/s] or [+7.491%; +7.904%]

scenario:line instrumentation - untargeted

• 🟩 throughput [+5477.250op/s; +5579.892op/s] or [+10.152%; +10.342%]

scenario:method instrumentation

• 🟩 throughput [+12599.350op/s; +13185.736op/s] or [+7.714%; +8.073%]

scenario:tracing - Propagation - Trace Context

• 🟥 throughput [-4100.162op/s; -4001.702op/s] or [-10.970%; -10.706%]

[ivoanjo]

I did a pass on the profiling bits and looks reasonable!

[Strech]

I did a pass over AppSec and can 👍🏼

BUT I do think that it makes sense to use AI to give a pass over some standards like top-level class instead of potential redefined class (::String instead of String)