[Feature Request] Find appropriate memory/cpu requirements for rosbridge service

Currently the rosbridge service uses the same memory/cpu constraints in the helm chart as the data service. But since the rosbridge service is not required to keep the entire file contents in memory, it's memory requirements could be reduced.

Profile the rosbridge service to find a "good enough" default memory/cpu configuration that would allow scaling multiple rosbridge pods on the same cluster but not impact performance too dramatically.

Start with the Helm chart's rosbridge resource constraints and identify how the service is deployed. Profile rosbridge under representative workloads, compare resource use and performance while scaling pods, and document defaults that reduce resource usage without a dramatic performance impact.

Written by the indexing model from the issue text.