Why We Needed a Smarter Way to Run CloudOps?
If you ever had the chance to work in Cloud operations, you must be familiar with the drill. Alert dashboards are overflowing, runbooks are buried in wikis, and hours are wasted simply stitching together the failure points before a solution can even be found. The sheer reality is that most of the time, operations teams end up reacting instead of staying ahead.
AIOps, or Artificial Intelligence for Operations, uses AI to cut through that noise by automating monitoring, identifying issues, and accelerating resolution.
Fueled by a determination to outsmart operational chaos, we found our answer in a powerful trio: Amazon Q Developer CLI, Custom Agents, and MCP Servers.
Together, these three give us a practical way to realize AIOps in the real world, following AWS’s vision of Observe → Engage → Act.
This blog explores how to modernize CloudOps using Amazon Q Developer CLI and MCP Servers, with Custom Agents, reducing the manual load. The outcome is simple: AIOps solutions that improve efficiency, deliver deeper insights, and scale with your operations.
Amazon Q Developer CLI: The Cloud Ops Teammate You Can Rely On
The Amazon Q Developer CLI is not just a chatbot bolted onto the terminal. Instead, it is your new agentic star of the team. It can read and write files, run shell commands (including AWS CLI), and perform multi-turn conversations that ensure actual execution. Above all, it explains the thought process and guides onto the next logical action item.
For example, instead of just telling you to run a command like npm install or aws ec2 describe instances, Q Developer can execute it directly after your confirmation. Imagine asking:
q chat "Which EC2 instances crossed 80% CPU in the last hour and why?"
q chat "Propose two remediations and apply the safer one after my approval."This is the type of fluid, conversational experience in operations that minimizes the mean-time-to-resolution (MMTR) drastically. The CLI also accelerates development by understanding codebases and guiding actions through adaptive dialogue.
Cloudelligent is already helping customers integrate this enhanced CLI as a vital part of their operational workflows. But to truly scale this power and manage complex environments, single agents hit a wall, that’s the domain of Custom Agents. Let’s dive into that next.
How We Keep Q Developer CLI Focused: Custom Agents
One of the biggest productivity killers is context switching. You might be debugging a backend API one moment and tagging a production alert the next. A single AI assistant can struggle in these situations, with overlapping tools or even confusing contexts. This is where Custom Agents come in to save your day.
Custom Agents act as dedicated workspaces optimized for your specific workflows. Each agent is configured with the right tools, permissions, and context for the task at hand. Switching between agents is seamless, so Q Developer CLI always knows which context and tools to use.
For example, you might employ:
- A front-end agent that focuses on React and Figma workflows.
- A back-end agent that handles Python and PostgreSQL.
- An ops agent that can read runbooks and use the AWS CLI, but won’t perform destructive actions without your approval.
You can control visibility to external systems and specify which files should be read by default. You can also include dynamic hooks, such as loading the current git status at the start of a session. These guardrails follow the principle of least privilege, keeping your workflow seamless while ensuring Q Developer CLI stays both powerful and safe.
Feeding Q Developer CLI the Context It Was Missing: MCP Servers
Even the smartest AI assistant will end up guessing without proper context. Model Context Protocol (MCP) servers bridge this gap by integrating Q Developer CLI directly into live systems, databases, logs, monitoring tools, or even knowledge bases. In practice, this means Q Developer CLI can counter-check your PostgreSQL schema before writing SQL or even pull Amazon CloudWatch metrics before diagnosing a performance issue.
For example, a PostgreSQL MCP configuration may look like this:
{
"mcpServers": {
"postgres-ro": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-postgres",
"postgresql://USER:PASS@HOST:5432/DBNAME"]
}
}
} If we ask ‘list slow queries’ now, Q Developer CLI will not hallucinate but query the schema, interpret the data and generate precise results. The flowchart below shows how the MCP Client API connects Amazon Q Developer CLI with external systems like Amazon Bedrock Knowledge Base and AWS Documentation. This enriches Q’s capability to make informed decisions and automate tasks.
Figure 1: MCP Client API integration enabling context enrichment for Amazon Q Developer CLI
Our AIOps Blueprint: How It All Fits Together
So how do all these pieces make the perfect jigsaw? Amazon Q Developer CLI, Custom Agents, and MCP servers combine into AIOps. This only works if they are supported and mapped to AWS’s simple model: Observe, Engage, and Act.
Figure 2: AIOps Blueprint: Observe → Engage → Act
It all starts with MCP servers feeding Q Developer CLI real-time context from logs, metrics, and schemas. Then, the CLI engages operators in conversations and proposes fixes backed by runbook knowledge. And finally, it acts by executing remediations safely with confirmation and audit trails in place.
flowchart LR
U[Engineer] -->|Natural language| QCLI[Q Developer CLI (agent)]
QCLI --> AG[Custom Agent (ops-agent)]
AG -->|context| MCP[MCP Servers]
MCP -->|metrics/logs/docs| CTX[Context Enrichment]
QCLI -->|proposal + ask-to-run| U
U -->|approve| QCLI
QCLI -->|safe actions| AWS[AWS APIs/CLI] We use this blueprint as the foundation of AIOps workshops where we help teams experience the end-to-end flow live.
Let’s See It in Action: Three Real Ops Scenarios
Let’s take a look at three real operational use cases that can be automated with Q Developer CLI and MCP.
High CPU on EC2: Q Developer CLI identifies spikes from Amazon CloudWatch and explains likely root causes and proposes remediations. With approval, it can scale an instance, restart a service, or roll back a deployment.
Figure 3: High CPU on EC2
Public Amazon S3 Bucket: Q Developer CLI lists down all buckets, flags public access, and walks you through the blast radius. With confirmation, it applies AWS’s block-public access settings and fixes the policy.
Figure 4: Public Amazon S3 Bucket
Unwanted open port: Q Developer CLI inspects the security groups and identifies an exposed port (0.0.0.0/0:5080) and recommends the least-privilege fix. With approval, it closes the port and updates the safety rules.
Figure 5: Unwanted Open Port
These are not hypothetical demos but are real working scenarios that ensure quick wins and show immediate ROI in reduced risk and downtime.
What We Learned Along the Way
The gains are undeniable. You can experience faster detection, consistent remediation, and shared configs that keep the teams aligned. But we come down to lessons that are just as important:
- Over-trusting tools will always be risky. If you allow Q Developer CLI to run destructive commands without approval, you will risk unintended changes.
- Knowledge bases must stay up to date, or you risk automated fixes that are outdated.
- Without audit trails, there is a risk of losing compliance visibility.
At Cloudelligent, we work with customers to design proper approval workflows, align runbooks with compliance standards, and set up regular KB refreshes. We humans remain in the loop, even as AI accelerates your workflow.
Go from Reactive to Proactive with Cloudelligent’s AIOps Blueprint
The potential here is estimated to be well beyond the use cases. MCP servers enable you to extend the Amazon Q Developer CLI to multi-cloud or hybrid environments. You can tie it into ServiceNow, Jira, or PagerDuty and ensure remediation proposals appear in your existing workflows. You can also create specialized agents for security, compliance, or cost optimization that follow a similar framework.
This approach lays the foundation for true AIOps. Self-healing systems can identify, diagnose, and remediate issues automatically, with operators stepping in only for exceptions.
Cloudelligent’s secret to building AIOps is simple: the right combination of Q Developer’s enhanced CLI for execution, Custom Agents for scoped focus, and MCP servers for context. Together, they transform operations from reactive firefighting to intelligent, proactive problem-solving.
If you’re ready to stop juggling dashboards and start building real AIOps, we’d love to help.
Talk to an expert at Cloudelligent today about leveraging our AIOps blueprint to transform your operations. Our experts will guide you through building secure agent profiles, integrating MCP servers, and deploying knowledge bases. Your operations team can then sit back and achieve faster, safer, and smarter outcomes.
