Investigate increased mem overhead in cluster host mem test

There has been a flake here ...

```console
[ RUN ] IpVersions/ClusterMemoryTestRunner.MemoryLargeHostSizeWithStats/IPv4
test/integration/stats_integration_test.cc:451: Failure
Expected: (m_per_host) <= (3005), actual: 3239 vs 3005
Stack trace:
0xb01a6f: Envoy::(anonymous namespace)::ClusterMemoryTestRunner_MemoryLargeHostSizeWithStats_Test::TestBody()
0x1e79d6c: testing::internal::HandleExceptionsInMethodIfSupported<>()
0x1e79c2d: testing::Test::Run()
0x1e7ae47: testing::TestInfo::Run()
... Google Test internal frames ...

```

https://dev.azure.com/cncf/envoy/_build/results?buildId=144930&view=logs&j=8c169225-0ae8-53bd-947f-07cb81846cb5&t=4d0399a5-fd0d-5446-d022-32daa00183cc&l=528

... which we will initially workaround - this ticket is to track further investigation