ITSM OPS — Effective ADLC Usage

8-step ITSM ticket lifecycle (intake → discover → classify → cross-validate → decide → implement → operate → govern) executed via cloudops/.adlc submodule consumption with PII-gated AWSO/ notebooks and real-time JIRA OPS board sync.

Project Identity

Aspect	Value
Repository	/Volumes/Working/projects/cloudops
JIRA Board	OPS (https://1xops.atlassian.net/jira/software/c/projects/OPS/)
Confluence Space	OPS (https://1xops.atlassian.net/wiki/spaces/OPS)
Sub-Product	CloudOps (CO-* story prefixes)
Data Boundary	AWSO/ folder (gitignored; customer account IDs, PII, $)

Consumption Pattern

The cloudops project symlinks to .adlc framework:

• .claude → .adlc/.claude (all 38 agents, 141 commands, 128 skills inherited)
• Local AWSO/ notebooks extend with customer-specific queries
• JIRA OPS board = source-of-truth for ticket lifecycle
• Confluence OPS = runbook + knowledge base for on-call teams

Daily Workflow: ITSM Ticket Lifecycle

Every OPS ticket flows through 8-step pipeline (see golden path docs/docs/golden-paths/itsm-lifecycle.md):

1. Intake → sre-engineer extracts: service type, environment, accounts, resource IDs
2. Discover → Multi-account infrastructure inventory (Cost Explorer, Config Aggregator, CloudTrail, SSM)
3. Classify → JIRA label taxonomy: 6-prefix (URGENCY, SERVICE, ENVIRONMENT, IMPACT, PRIORITY, RESOLUTION)
4. Cross-Validate → 4-way verification: API vs Console vs Config Agg vs CLI READONLY profiles
5. Decide → Change scheduling eligibility, blast radius, CAB routing
6. Implement → Create change record (CR) with change request number
7. Operate → Configure monitoring, alarms, escalation per Change Advisory Board decision
8. Govern → Cost impact summary, compliance evidence trail

P0/P1 Only: If incident severity ≥ P0, generate Post-Incident Review (PIR) → potential product story (SPM) if pattern recurs.

Effective Commands (This Project)

Command	Purpose	Input	Output
/itsm:lifecycle OPS-NNN	Full 8-step ticket processing with preview	JIRA ticket key	JIRA description updated with findings
/itsm:classify OPS-NNN	Auto-assign labels + priority + service type	Ticket description	Label recommendations in JIRA
/itsm:cross-validate OPS-NNN	4-way accuracy check (99.5% target)	Ticket scope	Cross-validation report in Confluence
/itsm:create-change OPS-NNN	Generate change record for CAB approval	Classification + decision	JIRA change request (subtask)
/itsm:create-pir OPS-NNN	Blameless post-incident review (P0/P1)	Resolved ticket + timeline	PIR document in Confluence + product story candidate
/sync:jira-pull	Fetch latest OPS tickets (batch)	JIRA JQL filter	Local CSV cache for analysis
/sync:jira-push	Sync local ticket changes → JIRA	Evidence JSON + markdown	JIRA ticket updated, activity logged
/documentation:confluence-publish	Publish runbooks + analysis → Confluence OPS space	Markdown source file	Live Confluence page with metadata

All commands default to preview mode. Add --execute to apply changes to JIRA + Confluence.

2-Way Sync: JIRA ↔ Local SSOT

Direction	Cadence	Mechanism	SSOT
JIRA → Local	Daily standup (morning)	/sync:jira-pull with JIRA Query Language filter	Live JIRA board state
Local → JIRA	Per-ticket resolution	/sync:jira-push with rsid label idempotency	Local evidence files (evidence/)
OPS → SPM	Escalation on pattern recurs	PIR tool flags product stories	product-owner agent reviews + stories.csv

Idempotency: Every JIRA sync operation includes rsid:{uuid} label. Running /sync:jira-push twice on same evidence = idempotent (no duplicates).

PII Boundary (Critical)

• Gitignored: AWSO/ folder contains customer names, account IDs, cost figures, sensitive runbooks
• Not tracked: Customer-specific Jupyter notebooks, live environment secrets, billing data
• Tracked: Framework logic (Python), patterns (lessons learned), generic runbooks (without account IDs)
• Publishing rule: All Confluence pages auto-scrub account IDs → replace with {account-id}, {region} placeholders

CxO Personas

Each ITSM output targets specific stakeholder:

Persona	Needs	Output Format
CFO (Cost)	Cost of incident (downtime × rate) + mitigation spend	Executive summary: 3 bullets + 1 $ impact
CTO (Architecture)	Root cause + fix duration + preventive design change	Technical runbook with architecture diagram
CloudOps Engineer (On-Call)	Step-by-step remediation + escalation matrix	Runbook.md with Rich CLI output examples
Customer Success	Timeline of impact + communication + resolution proof	Incident closure report with timestamps

Effective Hooks (ADLC Enforcement)

Hook	Trigger	Blocks	Purpose
enforce-coordination.sh	PreToolUse (ITSM commands)	Requires PO+CA logs	Prevents standalone execution on customer tickets
validate-bash.sh	Bash calls with AWS API	Blocks mutation verbs (delete, terminate) without HITL	READONLY only for autonomous team execution
validate-component-counts.sh	End of session	Audit trail	Verifies ITSM + SPM board state consistency

Quality Gates (ITSM-Specific)

1. Real-Time MCP Validation — Every JIRA ticket updated via MCP atlassian-tools (exact name: atlassian-tools>=0.21.0)
2. INVEST Adherence — Escalated product stories must meet INVEST criteria (stories.csv only)
3. Change Request Routing — CR blast radius determines CAB escalation + approval level
4. 4-Way Cross-Validation — READONLY AWS profile + Config Aggregator + Cost Explorer + Console output must agree ≥99.5%

Anti-Patterns Specific to ITSM

Anti-Pattern	Example	Prevention
CONFLUENCE_DUPLICATE_PAGE	PIR runbook created twice without title search	Hook blocks on title search before create
MCP_ADF_FIDELITY_LOSS	Rich taskList lost in JIRA description (MCP markdown→ADF)	Use REST API v3 for interactive checklists
JIRA_HIERARCHY_ASSUMPTION	Attaching PIR as child to resolved ticket, assuming hierarchy	Query parent type first; only Epics + Stories support children
JSON_ONLY_NATO	Evidence JSON updated without updating Confluence runbook	Rules-layer: update both in same session

Real Software Deliverables (2026-2030)

• Per-Customer Jupyter Notebooks (AWSO/) — Interactive analysis, READONLY cloud queries
• Inventory CSVs (AWSO/data/) — Account mappings, resource indices, historical snapshots
• Confluence Runbooks (OPS space) — Generic + customer-specific playbooks, PIRs
• JIRA Tickets (OPS board) — Live incident tracking, CR routing, resolution evidence
• Sync Evidence (evidence/) — JSON files for audit trails, traceability

References

• Golden Path: docs/docs/golden-paths/itsm-lifecycle.md (end-to-end flow, 8 steps)
• Framework: CLAUDE.md (root) → .adlc submodule configuration
• Anti-Patterns: .claude/rules/governance/anti-patterns-catalog.md (100 patterns, ITSM-specific callouts)
• JIRA Integration: .claude/rules/engineering/atlassian-integration-standards.md (API, MCP, sync patterns)
• READONLY AWS: .claude/rules/governance/operational-efficiency.md Rule 8 (autonomous execution with provided profiles)