openshift-eng RDS Analyzer workflow skill
Use this skill when guiding users through the end-to-end workflow from cluster data to deviation reports and optional Jira follow-up. It mirrors rds-analyzer docs/full-workflow.md and matches the current openshift-kni/rds-analyzer CLI (internal/cli/root.go, README.md).
Prerequisites (workflow context)
- Must-gather (or live cluster access via kubeconfig) — see Red Hat docs for
oc adm must-gather. - cluster-compare (kube-compare) plugin — produces the JSON input RDS Analyzer consumes. Install per OCP docs and upstream kube-compare.
- Telco reference — pick RAN DU, Telco Core, or Telco Hub; run comparison per product docs so output is JSON (e.g.
--output jsonwhere applicable).
Upstream rds-analyzer CLI (single command)
The repo exposes one executable with no subcommands: rds-analyzer [flags]. Cobra also provides --help / -h and --version / -v.
| Flag | Short | Default | Purpose |
|---|---|---|---|
--input |
-i |
stdin if not set | Path to kube-compare JSON. Without -i, data must be piped to stdin; interactive stdin with no pipe fails fast. |
--rules |
-r |
./rules.yaml |
Path to one rules YAML file. |
--output |
-o |
text |
text or html only. |
--output-mode |
-m |
simple |
simple or reporting. |
--target |
-t |
(empty) | OCP version for rules (e.g. 4.19). If empty, highest version in rules is used. |
--validate-rules-only |
— | false |
Validate regexp patterns in rules; no JSON input; do not use with -i. |
Examples (from upstream README / CLI help)
cat results.json | rds-analyzer -r /path/to/custom-rules.yaml
rds-analyzer -r ran-du-rules.yaml --validate-rules-only
rds-analyzer -i results.json -o html > report.html
rds-analyzer -i results.json -t 4.19
rds-analyzer -i results.json -r /path/to/rules.yaml
rds-analyzer -i results.json -m reporting
rds-analyzer --help
rds-analyzer --version
Container (README)
podman run --rm -v $(pwd):/data:Z quay.io/rhsysdeseng/rds-analyzer:latest \
-i /data/results.json -r /data/rules.yaml
Workflow scenarios (map to user requests)
A. Text report (terminal)
rds-analyzer -i comparison-results.json -r example-ran-du-rules.yaml
Equivalent defaults: -o text -m simple.
B. HTML report
rds-analyzer -i comparison-results.json -r example-ran-du-rules.yaml -o html > deviation-report.html
C. Specific OCP version + rules path
rds-analyzer -i comparison-results.json -o html -t 4.21 -r /path/to/rules.yaml > deviation-report.html
D. Reporting mode (LLM / Jira prep)
rds-analyzer -i comparison-results.json -r example-ran-du-rules.yaml -m reporting
E. Stdin
cat comparison-results.json | rds-analyzer -r rules.yaml
F. Validate rules only
rds-analyzer -r ran-du-rules.yaml --validate-rules-only
Helper script (optional)
From the ai-helpers repo, after chmod +x:
plugins/rds-analyzer/scripts/run_rds_analyzer.sh
text→-o text -m simplehtml→-o html -m simplereporting→-o text -m reportingvalidate-rules→--validate-rules-only(matches upstream flag name)
Example:
export RDS_ANALYZER_BIN=/path/to/rds-analyzer
./plugins/rds-analyzer/scripts/run_rds_analyzer.sh text -- -i results.json -r rules.yaml
./plugins/rds-analyzer/scripts/run_rds_analyzer.sh validate-rules -- -r rules.yaml
Reference documentation links (cluster compare / telco)
After the report: Jiras
Use deviation output to open or track issues. In ai-helpers, the jira plugin can assist with creating or structuring issues from structured text; combine reporting mode output with /jira:create or related commands when deviations require tracking or resolution.
Example rules path in rds-analyzer repo
examples/example-ran-du-rules.yaml (relative to a clone of openshift-kni/rds-analyzer).
Rules engine reminders
- Worst impact wins: Impacting > NeedsReview > NotImpacting > NotADeviation.
- Regex in rules is validated at startup; invalid patterns fail before JSON is read (unless
--validate-rules-onlyalone).