[MAIN-2169] Prometheus preformance fix

[Avi-Robusta]

We updated prometrix for faster preformance.
https://pypi.org/project/prometrix/0.2.0/

There were issues with speed and cpu when serialising the Hikaru object.
you can see the speed improvements here

As you can see it is used practically in the same way but now needs some slight changes

• dict(block.data) to de-serialise it
• the sub objects from PrometheusQueryResult are all treated as a dict object
  • i.e instead of query_result.vector_result[0].value we need to use query_result.vector_result[0]["value"]

Tests ran

• UI - (krr, nodes, metrics exporter, apps page + apps custom graphs)
• DB - (verified the node average usage query reports correct data to DB)
• Actions - actions impacted - prometheus_sla_enricher - graph enrichers like node_graph_enricher

[arikalon1]

nice work

left some comments

[Avi-Robusta]

Clarifying the types here @arikalon1
https://github.com/robusta-dev/prometrix/pull/26/files

[arikalon1]

nice work