Skip to main content

SRE Skills (9)

Part of Role Skills Catalog | Phase 4 + Phase 6

Chains: SLO → Alerts | Pipeline Create → Pipeline Audit → Env | Toil → Hardening | Runbook → Postmortem

Userflow Schema​

flowchart TD
    jaan-to-sre-slo-setup["sre-slo-setup<br>SLO Setup<br>SLOs + error budgets + SLIs"] --> jaan-to-sre-alert-tuning["sre-alert-tuning<br>Alert Tuning<br>Noise reduction + severity map"]
    jaan-to-sre-alert-tuning["sre-alert-tuning<br>Alert Tuning<br>Noise reduction + severity map"] -.-> jaan-to-dev-observability-alerts["dev-observability-alerts<br>DEV: observability-alerts"]
    detect-dev["detect-dev<br>DEV: stack-detect"] -.-> jaan-to-sre-pipeline-create["sre-pipeline-create<br>Pipeline Create<br>CI/CD workflows + Docker + deploy"]
    jaan-to-sre-pipeline-create["sre-pipeline-create<br>Pipeline Create<br>CI/CD workflows + Docker + deploy"] --> jaan-to-sre-pipeline-audit["sre-pipeline-audit<br>Pipeline Audit<br>Weak points + missing gates"]
    jaan-to-sre-pipeline-audit["sre-pipeline-audit<br>Pipeline Audit<br>Weak points + missing gates"] --> jaan-to-sre-env-check["sre-env-check<br>Env Check<br>Config drift + readiness"]
    jaan-to-sre-env-check["sre-env-check<br>Env Check<br>Config drift + readiness"] -.-> jaan-to-dev-ship-check["dev-ship-check<br>DEV: ship-check"]
    jaan-to-sre-toil-audit["sre-toil-audit<br>Toil Audit<br>Toil list + automation candidates"] --> jaan-to-sre-hardening-plan["sre-hardening-plan<br>Hardening Plan<br>Reliability backlog + steps"]
    jaan-to-sre-hardening-plan["sre-hardening-plan<br>Hardening Plan<br>Reliability backlog + steps"] -.-> jaan-to-delivery-backlog-split["delivery-backlog-split<br>DELIVERY: backlog-split"]
    jaan-to-sre-incident-runbook["sre-incident-runbook<br>Incident Runbook<br>Runbook + escalation + triage"] --> jaan-to-sre-incident-postmortem["sre-incident-postmortem<br>Incident Postmortem<br>Timeline + actions + owners"]
    jaan-to-sre-incident-postmortem["sre-incident-postmortem<br>Incident Postmortem<br>Timeline + actions + owners"] --> jaan-to-sre-hardening-plan["sre-hardening-plan<br>Hardening Plan<br>Reliability backlog + steps"]

    style detect-dev fill:#f0f0f0,stroke:#999
    style jaan-to-dev-observability-alerts fill:#f0f0f0,stroke:#999
    style jaan-to-dev-ship-check fill:#f0f0f0,stroke:#999
    style jaan-to-delivery-backlog-split fill:#f0f0f0,stroke:#999

Legend: Solid = internal | Dashed = cross-role exit | Gray nodes = other roles

/jaan-to-sre-slo-setup​

  • Logical: sre-slo-setup
  • Description: SLOs + error budgets, suggested SLIs (latency, errors, uptime), reporting cadence
  • Quick Win: No - needs monitoring data
  • Key Points:
    • Define SLOs and error budgets
    • Alerts should reflect user impact
    • Tune noise and prioritize
  • → Next: sre-alert-tuning
  • MCP Required: Monitoring tools (Datadog/Grafana)
  • Input: [service]
  • Output: $JAAN_OUTPUTS_DIR/sre/monitor/{slug}/slo-setup.md

/jaan-to-sre-alert-tuning​

  • Logical: sre-alert-tuning
  • Description: Noise reduction plan, priorities + severity mapping, missing alerts checklist
  • Quick Win: No - needs alert history
  • Key Points:
    • Define SLOs and error budgets
    • Alerts should reflect user impact
    • Tune noise and prioritize
  • → Next: dev-observability-alerts
  • MCP Required: Monitoring tools (Datadog/Grafana)
  • Input: [alerts]
  • Output: $JAAN_OUTPUTS_DIR/sre/monitor/{slug}/alert-tuning.md

/jaan-to:sre-pipeline-create​

  • Logical: sre-pipeline-create
  • Description: Generate CI/CD pipeline configuration: GitHub Actions workflows, Dockerfile, deployment manifests, testing stages
  • Quick Win: Yes
  • Key Points:
    • Multi-stage Docker builds for optimization
    • Parallel test/lint/type-check jobs
    • Environment-specific deployment with approval gates
    • Secrets management pattern (GitHub Secrets)
  • → Next: sre-pipeline-audit
  • MCP Required: None
  • Input: [tech-stack, project-structure, deployment-target]
  • Output: $JAAN_OUTPUTS_DIR/sre/cicd/{slug}/pipeline/
  • Note: Complements existing sre-pipeline-audit (audit-only)

/jaan-to-sre-pipeline-audit​

  • Logical: sre-pipeline-audit
  • Description: Weak points + quick fixes, missing gates checklist, reliability improvements backlog
  • Quick Win: No - needs GitLab MCP
  • Key Points:
    • Identify weak gates and flaky tests
    • Ensure env parity and secret management
    • Make rollbacks routine
  • → Next: sre-env-check
  • MCP Required: GitLab (pipeline config)
  • Input: [repo]
  • Output: $JAAN_OUTPUTS_DIR/sre/cicd/{slug}/pipeline-audit.md

/jaan-to-sre-env-check​

  • Logical: sre-env-check
  • Description: Config drift + readiness checklist, missing secrets/configs, "safe to deploy?" hints
  • Quick Win: No - needs GitLab MCP
  • Key Points:
    • Identify weak gates and flaky tests
    • Ensure env parity and secret management
    • Make rollbacks routine
  • → Next: dev-ship-check
  • MCP Required: GitLab (env config)
  • Input: [env]
  • Output: $JAAN_OUTPUTS_DIR/sre/cicd/{slug}/env-check.md

/jaan-to-sre-toil-audit​

  • Logical: sre-toil-audit
  • Description: Toil list + automation candidates, time spent estimates, top ROI opportunities
  • Quick Win: Yes
  • Key Points:
    • Track toil weekly; automate repeatable work
    • Maintain reliability backlog
    • Prioritize hardening with measurable outcomes
  • → Next: sre-hardening-plan
  • MCP Required: None
  • Input: [week]
  • Output: $JAAN_OUTPUTS_DIR/sre/ops/{slug}/toil-audit.md

/jaan-to-sre-hardening-plan​

  • Logical: sre-hardening-plan
  • Description: Reliability improvement backlog, sequenced steps, verification metrics
  • Quick Win: Yes
  • Key Points:
    • Track toil weekly; automate repeatable work
    • Maintain reliability backlog
    • Prioritize hardening with measurable outcomes
  • → Next: delivery-backlog-split
  • MCP Required: None
  • Input: [service]
  • Output: $JAAN_OUTPUTS_DIR/sre/ops/{slug}/hardening-plan.md

/jaan-to-sre-incident-runbook​

  • Logical: sre-incident-runbook
  • Description: Runbook + escalation steps, triage checklist, "if X then Y" actions
  • Quick Win: Yes
  • Key Points:
    • Runbooks should be actionable and short
    • Clear escalation paths
    • Postmortems must yield backlog items
  • → Next: sre-incident-postmortem
  • MCP Required: None
  • Input: [service]
  • Output: $JAAN_OUTPUTS_DIR/sre/incident/{slug}/runbook.md

/jaan-to-sre-incident-postmortem​

  • Logical: sre-incident-postmortem
  • Description: Blameless retro template filled, timeline + contributing factors, action items + owners
  • Quick Win: Yes
  • Key Points:
    • Runbooks should be actionable and short
    • Clear escalation paths
    • Postmortems must yield backlog items
  • → Next: sre-hardening-plan
  • MCP Required: None
  • Input: [incident]
  • Output: $JAAN_OUTPUTS_DIR/sre/incident/{slug}/postmortem.md