Automate Infrastructure Maintenance and Optimization

This agent proactively monitors system health, optimizes resource allocation, manages scaling decisions, and maintains infrastructure reliability with minimal human intervention. It starts with health assessment: analyzing system metrics (CPU, memory, disk, network), checking service availability and response times, reviewing error logs for patterns, and validating backup and disaster-recovery readiness.

Performance analysis evaluates resource utilization trends over time, identifies bottlenecks and degradation, compares current metrics against baselines, and assesses capacity-planning needs. Scaling decisions determine when horizontal or vertical scaling is needed, calculate optimal resource allocation, evaluate cost implications, and prioritize by business impact. Maintenance planning schedules maintenance windows, plans security updates and patches, coordinates dependency updates, and prepares rollback strategies. Issue resolution diagnoses infrastructure problems, implements automated fixes where safe, escalates critical issues with detailed context, and documents solutions.

It applies fixed decision thresholds: scale up when CPU exceeds 80% for 15+ minutes, memory exceeds 85%, or response time exceeds 2x baseline; scale down when CPU is under 30% for 2+ hours with memory under 50% and stable traffic; alert critical on any service downtime, error rate above 5%, or disk above 90%; and schedule maintenance with security patches within 48 hours and other updates within a week. Automation triggers cover auto-scaling for predictable traffic patterns, auto-restart for known memory-leak/recovery patterns, auto-backup before major changes, and auto-rollback if error rates increase more than 10x after a deployment.

The output is an Infrastructure Status Report (overall health status, critical issue count, system metrics with trends, per-service status table, scaling recommendations with cost impact, maintenance schedule, and prioritized action items) and, when scaling is being considered, a Scaling Decision Matrix comparing vertical vs. horizontal scaling options with cost impact, expected performance gain, a recommendation with justification, an implementation timeline, and a rollback plan. Guidelines throughout: proactive monitoring to anticipate issues, cost-optimized decisions, safety-first changes with rollback plans, detailed documentation, automation preferred over manual intervention, and clear stakeholder communication.

Use this agent for ongoing infrastructure monitoring, scaling decisions, and maintenance planning where cost, performance, and risk all need to be weighed together - especially for services with defined utilization thresholds to act on. It is well suited to teams running production infrastructure that needs proactive rather than reactive management. It is not meant for one-off infrastructure provisioning with no ongoing monitoring need, or for systems too small to have meaningful scaling decisions to make.

Input: system metrics (CPU, memory, disk, network), service status, and error logs.

Output: an Infrastructure Status Report and, when relevant, a Scaling Decision Matrix. Example status report structure:

# Infrastructure Status Report - [Date]

• Overall system health: [GREEN/YELLOW/RED]
• Critical issues: [Number]
• Recommended actions: [Number]

• CPU utilization: [%]
• Memory usage: [%]

Analyzes system metrics, service-availability data, and error logs from existing monitoring infrastructure; it does not connect to a specific monitoring or cloud platform itself.

SRE and infrastructure teams needing proactive health monitoring and scaling decisions, and teams that want cost and risk projections attached to every scaling and maintenance recommendation.