Below are the existing network microbenchmarks:

• b_first_half
  • observable_mutex_bm
• b_second_half
  • session_workflow_bm
  • service_executor_bm

Currently, microbenchmarks outside benchmarks_sep task are not tracked and do not automatic perf BF generation. We should introduce benchmarks_network task (similar to benchmarks_query) and define a Performance Triage Context for it similar to as described in PERF-9148. 

Relatedly, the network microbenchmarks don't use runBenchmarkWithProfiler (e.g. in. the service_entry_point_shard/router_role_bm. So they only have the latency_per_op* metrics instead of also cycles_per_iteration* and instructions_per_op*. We should considering adding profiling to those microbenchmarks.