Change subsection title

Author: rhshadrach

web/pandas/community/benchmarks.md

@@ -24,7 +24,7 @@ running in the same hardware with different levels of stress have a big impact i
 the result. Even running the benchmarks with identical hardware and almost identical
 conditions can produce significant differences when running the same exact code.
 
-## pandas benchmarks servers
+## Automated benchmarks runners
 
 We currently have two setups running the benchmarks of pandas for every
 (or almost every) commit to the `main` branch. One is run on GitHub actions