1. AI DevOps Platform

Submit AI HelpDesk tickets in the DuploCloud Portal

You can use DuploCloud's DevOps Agent to troubleshoot issues with your infrastructure, get guidance on common operational tasks, and receive suggested commands for Services, Containers, and cloud resources — all from within the DuploCloud Portal.

Using the AI HelpDesk

Quickly interact with the DevOps Agent to get guidance, suggestions, or commands for common operational tasks.

Navigate to AI DevOps → HelpDesk in the DuploCloud Portal and click Add Ticket. Select the Scope that gives the agent access to the infrastructure relevant to your task, and choose the Skills that define how the agent behaves — for example, Terraform skills contain instructions for IaC workflows, Kubernetes skills cover cluster operations, and FinOps skills focus on spend analysis. Then enter your request in the message field and click Submit.

The agent streams its response in real time. As it works, it may propose commands for you to Approve, Reject, Ignore, or Execute — no action is applied to your infrastructure without your sign-off.

Daily Use Cases

The following are common operational tasks to try in the HelpDesk. For each, select a scope that covers the relevant environment and include the skills that match the task domain.

Drift in IaC

• Scope: A scope with access to your Git repository and AWS account

• Skills: Terraform or IaC (guides the agent through infrastructure state management, drift detection, and remediation workflows)

• Message to Agent:

  Review our Terraform state and compare it against what's currently provisioned
  in AWS. Identify any resources that exist in AWS but are not tracked in state,
  or resources in state that no longer exist. Summarise the drift and suggest
  remediation steps.

Kubernetes Pod Health

• Scope: A scope with access to your Kubernetes cluster

• Skills: Site Reliability or Kubernetes (equips the agent with cluster operations, uptime analysis, and pod-level diagnostics)

• Message to Agent:

  Check the health of all pods across our Kubernetes cluster. Identify any pods
  in CrashLoopBackOff, Pending, or OOMKilled state. For each unhealthy pod,
  provide the relevant logs and a suggested fix.

Recent Alerts on Observability Tools

• Scope: A scope that includes your observability provider or a configured MCP server for that tool

• Skills: Site Reliability (focuses the agent on alert triage, severity analysis, and observability platform queries)

• Message to Agent:

  Pull the most recent alerts from our observability platform from the last
  24 hours. Group them by severity and service. Highlight any that are still
  active and suggest next steps for the top three by impact.

AWS Cost and Optimisation Suggestions

• Scope: A scope with read access to your AWS account

• Skills: FinOps or cost optimisation (directs the agent toward cloud spend analysis, resource rightsizing, and savings recommendations)

• Message to Agent:

  Analyse our AWS environment for cost inefficiencies. Look for oversized or
  idle EC2 instances, unattached EBS volumes, unused Elastic IPs, and any other
  common cost drivers. Provide a ranked list of recommendations with estimated
  monthly savings for each.