Investigate Strider connection errors on NodeNorm CI

[gaurav]

Aragorn/Strider is reporting a large number of connection timeouts when sending lots of requests to NodeNorm (see OTEL trace on CI at https://translator-otel.ci.transltr.io/trace/53f3945e7f9a5b34a8a26f1793b99cbb). Investigation revealed that there were no error messages in NodeNorm Web logs (as far as we could tell), and the presence of connection errors suggests that this is happening in Kubernetes before it gets to the web pod (but our logs may also be incomplete, see #275). We also found one Redis database (db1) running at 99.9% memory usage -- we'll upgrade that (helxplatform/translator-devops#935) and see if that makes any difference.