feat(jenkins): parse test_metadata from build description

[fruch]

Summary

• Adds parse_test_metadata_from_description() utility that extracts SCT pipeline metadata (tier, test_type, duration_class, supported_backends) from Jenkins job descriptions
• Supports two formats: job description (markdown ### TestMetadata block) and build description (HTML <b>Metadata:</b> block)
• Adds JenkinsService.get_build_test_metadata() method to fetch and parse metadata for a given job/build
• Includes unit tests for the parser with real-world examples

Context

SCT PR scylladb/scylla-cluster-tests#14818 adds test_metadata to test-case YAML files and writes it to the Jenkins job description during job creation. The format in the job description is:

test: longevity_test.LongevityTest.test_custom_time | backend: aws | region: eu-west-1 | config: test-cases/longevity/longevity-10gb-3h.yaml

Basic longevity test running cassandra-stress write workload at QUORUM consistency
for ~4 hours on a 6-node single-DC cluster with SisyphusMonkey nemesis. Validates
cluster stability under moderate write load with continuous chaos operations.

### TestMetadata
tier: tier1
test_type: longevity
duration_class: short
supported_backends: ['aws', 'gce', 'azure']

The parser extracts the ### TestMetadata block (key: value lines) from this description.

A secondary HTML format is also supported for backward compatibility with currentBuild.description:

<b>Metadata:</b> tier=tier1 | type=longevity | duration=short | backends=['aws', 'gce']

Scope

This PR only adds reading and parsing. Storage/persistence of the metadata (on ArgusTest, SCTTestRun, or elsewhere) is intentionally left for a design discussion with the team.

Discussion Points

• Where should this metadata live? Options:
  • New columns on ArgusTest (job-level, updated during scan-jenkins)
  • New columns on SCTTestRun (per-run, extracted from sct_config)
  • Both (job-level defaults + per-run overrides)
• Should JenkinsMonitor.collect() call get_build_test_metadata() and store it during scans?
• Should the frontend expose filtering/grouping by tier/test_type?

[fruch]

This is the suggested "interface" for the test metadata, it's gonna be available on the job description

this PR is just ref to see where it should handled, it's not saving or doing anything with the metadata yet

[soyacz]

looks very fragile.
Why not using yaml/json/xml?