Optional CI task: Benchmarks using rally

It would be good to have benchmarks directly in the CI pipeline.

Requirements:

• Manually triggered, not required for build/release#
• Runs as a weekly automated job. Use CI runner tags to separate nightly builds from benchmark.
• Uses OpenSearch Benchmark: https://github.com/opensearch-project/opensearch-benchmark
• Runs a baseline OpenSearch with standard security plugin against OpenSearch with Eliatra Suite with a predefined configuration using a representative configuration (using DLS, FLS, FM, etc)
• Runs on the same test runner as the integration tests do
• For the beginning, it just uses the standard rally http_logs track: https://esrally.readthedocs.io/en/stable/race.html#list-tracks
• Outputs the result as plain ASCII standard rally result, both to stdout and to a downloadable artifact