Add job footprint as a svelte component

[moebiusband73]

Apart from the polar plot the aggregate performance of a job is not shown. Add a component that renders an overall job footprint including energy consumption. It might be required to extent the db schema. It has to be configurable which metrics are used.

[spacehamster87]

Some questions for preparation / discussion:

• Use footprint-component as ...
  • additional card above job metric plots, enabled / disabled by default?
  • alternative to polar plot / roofline / or both?
  • enabled / switched via button or similar
• In addition to all available metric averages, which additional info is useful?
  • E.g., energy consumption (duration * avg. wattage) is mentioned in opening post

[spacehamster87]

Branch 196_add_footprint

• Adds footprint component
  • Renders base metric averages as colored bars respective to configured threshold values
  • Rendered metrics dependent on job and cluster
  • CPU Metrics (Exclusive): cpu_load, flops_any, mem_used, mem_bw
  • CPU Metrics (Shared): cpu_load, flops_any, mem_used
  • GPU Metrics (Exclusive): cpu_load, flops_any, acc_utilization, mem_bw
  • GPU Metrics (Shared): cpu_load, flops_any, acc_utilization
  • Metric-Threshold configs for sublucsters are used when found
  • Metric-Thresholds are scaled for shared jobs
  • cpu_load maximum is number of allocated cores
  • aggregation:sum metric thresholds are scaled down based on number of allocated cores
• Component always rendered in job-view
• Component optional in job-list, activated in "Metrics" selector
  • Selection saved for User and Cluster
• Additionally, new fields are requestable from GraphQL, read from jobs' DB entry
  • ´flopsAnyAvg, memBwAvg, loadAvg´
• Metadata field message is rendered at the bottom of the component, if found (see #220 )
  • May contain pre-styled html and is rendered accordingly

[spacehamster87]

Merged