Proposal — lightweight change-communication process
Goal: Reduce surprises/incidents by making planned infra changes visible, evaluable, and measurable without heavy ceremony.
Process summary
- All changes affecting infra or services -> short Change Notice in a central system at least 48 hours before work (emergency exceptions documented).
- Use a single source: an existing ticketing/ITSM tool (Jira Service Management) or a lightweight calendar + Slack integration for notifications.
Required fields (CHANGE NOTICE)
- Title, Owner (name + pager), Team
- Scope (systems, regions, clusters, services impacted)
- Change type (config, deploy, upgrade, maintenance, emergency)
- Start/End window + rollback window
- Risk level (low/med/high) + brief rationale
- Pre-checks and post-checks (health endpoints, metrics to watch)
- Rollback plan and contact on-call
- Stakeholders notified (teams/SLAs)
- Link to runbook/automation scripts
Review points
- Auto triage: low-risk changes auto-approve; medium/high require 1 peer review from affected team within 24h and ops approval for high-risk.
- Ops/On-call review for schedule conflicts, capacity, maintenance windows.
- Final “GO/NO-GO” posted to notice 15 minutes before start.
Tooling
- Jira ticket template + custom fields; webhook to a “change-feed” channel in Slack and to an on-call rota (Opsgenie/PagerDuty).
- Dashboard: Kibana/Grafana board showing upcoming changes, conflicts, and risk distribution.
- Optional: Git-backed runbook links (Confluence or docs repo).
Measurement
- Compliance: percent of changes created ≥48h before start; percent with required fields completed.
- Effectiveness: number of incidents attributable to changes per quarter; mean time to detect after change; % of successful rollbacks.
- Targets: ≥90% timely notices, ≤X change-related incidents (baseline then reduce 30% in 6 months).
- Monitor via dashboards and monthly post-change review (blameless RCA on incidents).
Why this works
- Minimal friction (template + auto-notify) keeps adoption high.
- Risk-based reviews focus attention where it matters.
- Metrics close the loop and drive continuous improvement.
