/full-workflow

Name

rds-analyzer:full-workflow

Synopsis

/rds-analyzer:full-workflow [scenario]

Description

Orchestrates the complete RDS Analyzer workflow from openshift-kni/rds-analyzer (docs/full-workflow.md):

1. Prerequisites — Must-gather (or cluster access) and kube-compare JSON as input.
2. Cluster compare — User must have the cluster-compare (kube-compare) plugin installed; official install and telco reference docs.
3. Telco references — RAN DU, Telco Core, Telco Hub documentation links for running comparisons.
4. RDS Analyzer — Run the single upstream rds-analyzer binary with the flags documented below (text/HTML, --target, --output-mode reporting, stdin, --validate-rules-only).
5. Jira / LLM follow-up — Reporting-mode output for telco team interaction; suggest Jira plugin commands when the user wants tickets.

This command does not replace installing cluster-compare or gathering must-gather; it guides execution and runs rds-analyzer when the user provides paths and a binary (PATH or RDS_ANALYZER_BIN).

Upstream CLI — all commands and flags (openshift-kni/rds-analyzer)

The repository ships one Cobra root command: rds-analyzer (no subcommands). Standard Cobra additions apply:

Invocation	Purpose
rds-analyzer [flags]	Default: read kube-compare JSON (from -i or stdin), load rules, write deviation report to stdout.
rds-analyzer --help	Short and long help (synopsis, flag list, embedded examples). Short: -h.
rds-analyzer --version	Print build version string. Short: -v.

Analysis flags (from internal/cli/root.go and README.md):

Flag	Short	Default	Description
--input	-i	(unset → stdin)	Path to kube-compare JSON. If omitted, JSON must be piped on stdin; if stdin is a TTY with no pipe, the tool errors (no input provided; use -i flag or pipe JSON data).
--rules	-r	./rules.yaml	Path to one rules YAML file.
--output	-o	text	Output format: text or html only.
--output-mode	-m	simple	Output mode: simple or reporting (LLM-oriented two-section layout).
--target	-t	(empty)	OCP version for rules evaluation (e.g. 4.19, 4.21). If unset, the engine uses the highest version defined in the rules file.
--validate-rules-only	(none)	false	Load rules YAML, validate all regexp / value_regex patterns, print success or error, exit without reading JSON. Cannot be combined with -i.

Upstream embedded examples (from CLI Long help text):

# stdin + custom rules
cat results.json | rds-analyzer -r /path/to/custom-rules.yaml

# validate rules only (flag name is --validate-rules-only, not --validate-rules)
rds-analyzer -r ran-du-rules.yaml --validate-rules-only

# file input + HTML
rds-analyzer -i results.json -o html > report.html

# target OCP version (default rules path ./rules.yaml if -r omitted)
rds-analyzer -i results.json -t 4.19

# custom rules + input file
rds-analyzer -i results.json -r /path/to/rules.yaml

# reporting mode
rds-analyzer -i results.json -m reporting

README / workflow doc examples (same binary; paths illustrative):

rds-analyzer -i comparison-results.json -r example-ran-du-rules.yaml
rds-analyzer -i comparison-results.json -r example-ran-du-rules.yaml -o html > deviation-report.html
rds-analyzer -i comparison-results.json -o html -t 4.21 -r /path/to/rules.yaml > deviation-report.html
rds-analyzer -i comparison-results.json -r example-ran-du-rules.yaml -m reporting

Container (from upstream README; image quay.io/rhsysdeseng/rds-analyzer):

podman run --rm -v $(pwd):/data:Z quay.io/rhsysdeseng/rds-analyzer:latest \
  -i /data/results.json -r /data/rules.yaml
podman run --rm -v $(pwd):/data:Z quay.io/rhsysdeseng/rds-analyzer:latest \
  -i /data/results.json -r /data/rules.yaml -o html > report.html
cat results.json | podman run --rm -i -v $(pwd):/data:Z quay.io/rhsysdeseng/rds-analyzer:latest \
  -r /data/rules.yaml

Implementation

1. Load the skill plugins/rds-analyzer/skills/rds-analyzer-workflow/SKILL.md for extra context and links.
2. Interpret scenario (optional $1):
   • prerequisites — Must-gather and environment checklist only.
   • cluster-compare — Install links and how to obtain JSON for RDS Analyzer.
   • telco-refs — Telco RAN DU / Core / Hub doc links for cluster compare.
   • analyze-text — Text report: rds-analyzer -i … -r … (defaults -o text -m simple).
   • analyze-html — HTML: rds-analyzer -i … -r … -o html (redirect stdout to a file).
   • analyze-reporting — LLM-oriented: -m reporting (still text stream unless combined with -o html as appropriate).
   • analyze-versioned — Include -t <OCP> when targeting a specific OCP version, and -r path when using non-default rules.
   • validate-rules — Run rds-analyzer -r <rules.yaml> --validate-rules-only (no -i, no stdin JSON).
   • jira-followup — How to use reporting output with Jira workflows (reference /jira:* commands in ai-helpers).
   • Omitted or all — Walk through sections 1–5 in order; ask for missing inputs before running commands.
3. Execute when the user supplies concrete paths:
   • Prefer rds-analyzer on PATH, or RDS_ANALYZER_BIN.
   • Optional wrapper: plugins/rds-analyzer/scripts/run_rds_analyzer.sh (text, html, reporting, validate-rules → passes --validate-rules-only to the binary).
4. Rules file — Upstream accepts one -r path per invocation. For combined hub + DU rules, merge YAML offline or run separate passes as the workflow requires.
5. Never combine -i (or stdin JSON) with --validate-rules-only.

Return Value

• Guidance: Step-by-step narrative for the chosen scenario(s).
• Commands: Copy-pasteable rds-analyzer (or podman run …, or script) invocations with user-provided paths.
• On failure: Parse stderr; for regexp errors suggest --validate-rules-only or fixing rules YAML.

Examples

1. Full guided workflow

   /rds-analyzer:full-workflow
   

2. Validate rules only (upstream flag)

   /rds-analyzer:full-workflow validate-rules
   

3. HTML report scenario

   /rds-analyzer:full-workflow analyze-html
   

4. Reporting mode for LLM / Jira prep

   /rds-analyzer:full-workflow analyze-reporting
   

5. Show upstream help / version (when the user asks how to run the tool)

   rds-analyzer --help
   rds-analyzer --version
   

Arguments

• $1 scenario (optional): One of prerequisites, cluster-compare, telco-refs, analyze-text, analyze-html, analyze-reporting, analyze-versioned, validate-rules, jira-followup, or all. If omitted, default to all. The scenario validate-rules maps to --validate-rules-only on the real binary.

See Also

• Skill: plugins/rds-analyzer/skills/rds-analyzer-workflow/SKILL.md
• Upstream workflow: docs/full-workflow.md
• Upstream usage index: USAGE.md
• Script: plugins/rds-analyzer/scripts/run_rds_analyzer.sh