Jeeves
Infrastructure & Sysadmin
The agent that keeps the infrastructure standing.
Online
Jeeves owns the servers. Every Docker stack, every update pipeline, every deployment — Jeeves handles it with the kind of quiet competence that means you only hear about it when something actually needs your attention. Infrastructure problems tend to get solved before anyone notices there was one.
By the numbers
15+
Services Maintained
Auto
Update Pipelines
0
Unplanned Outages
Always
Watching
What it does
Container Stack Management
Dozens of containerized services are maintained, updated, and monitored continuously. Updates are applied on schedule, health is verified after each change, and rollback happens automatically if something breaks.
⏱ Hours of manual ops/week
Security Update Pipelines
Packages, images, and dependencies are kept current without manual intervention. Updates are staged, tested, and applied — critical patches are fast-tracked, routine updates are batched.
⏱ Eliminates update debt
Service Health Monitoring
Uptime, response times, and service availability are tracked continuously across the entire stack. Status changes trigger immediate alerts — up or down. No news is good news.
⏱ Instant incident awareness
Deployment Orchestration
New services and updates are deployed using consistent, repeatable processes. Configuration is version-controlled, deployments are logged, and rollback procedures are always ready.
⏱ Consistent, reliable deploys
Infrastructure Change Control
All significant infrastructure changes are reviewed and approved before execution. No surprises — changes are proposed, evaluated, and applied deliberately, not impulsively.
⏱ Prevents configuration drift
Incident Response
When something breaks, Jeeves investigates, diagnoses, and resolves — or escalates with a clear description of what happened, what was tried, and what needs a human decision.
⏱ Faster mean time to recovery
Local vs cloud AI
Infrastructure management is overwhelmingly local — scripts, automation, and lightweight local models handle the day-to-day. Cloud AI is only involved when genuine reasoning about a novel problem is needed.
🏠 Runs locally
- Container deployments & updatesPure scripting — no AI
- Health monitoring & alertsLocal scripts + Uptime Kuma
- Log analysis & triagellama3.1:8b on local GPU
- Crontab & PM2 change detectionllama3.1:8b diff analysis
- Security update pipelinesAutomated local tooling
- Service restarts & rollbacksPure scripting — no AI
☁️ Uses cloud
- Incident diagnosis & planningClaude Sonnet
- Infrastructure design decisionsClaude Sonnet
- Complex runbook generationClaude Sonnet
How it works
Continuously
Service health checks run across all monitored endpoints
Nightly
Update pipelines check for new versions and stage changes
On change
Deployments triggered, health verified, status reported
On incident
Automatic diagnosis and resolution attempt before escalation
On request
Infrastructure changes proposed, reviewed, then executed