This past week presented a significant learning opportunity, combining technical root-cause analysis with workplace communication and problem-solving skills. My main objective was stabilizing remote access for an employee utilizing Azure Virtual Desktop. While remote desktop deployments aim to provide seamless connectivity, real-world execution often reveals unforeseen issues and administrative access barriers. Navigating these constraints reinforced how essential self-advocacy and data-driven troubleshooting are in cloud administration.

The initial challenge during this incident was not technical, but organizational. Early on, the user experienced recurring session drops that interrupted their productivity. However, I initially lacked read permissions within the Azure Portal to view AVD logs and perform diagnostics. My supervisor initially expressed frustration, feeling our team was not researching the issue thoroughly enough. Recognizing that standard client-side troubleshooting would not suffice, I communicated the specific visibility gaps preventing deeper analysis of the issue and clarified the steps I took to follow the appropriate L1 troubleshooting protocols prior to escalation. After demonstrating what logs were required to diagnose the root cause, my supervisor was receptive and granted me the necessary read access to AVD Insights, allowing me to conduct a thorough investigation.

Once granted access, I was able to correlate event logs with exact timestamps of the disconnects. Rather than relying on guess work or rebooting the VM, I looked into the behaviors surrounding the failures. The host’s RDStack logs specifically recorded an OutOfMemory (32) event right at the point of disconnection. Approximately one second after this memory exhaustion occurred, Desktop Windows Manager crashed, which stopped the client session. The log evidence confirmed that the host VM had insufficient RAM allocated to maintain reliable desktop connection.

I gathered my findings and presented my supervisor with the data in a report to my supervisor, since I was only given read permissions the next step was escalating it to those with the permissions to make the necessary changes. My supervisor was able to upgrade the virtual machine to a higher-tier plan with additional memory capacity. Following the upgrade, the employee was able to re-establish their session and maintain uninterrupted connections throughout the rest of the week.

This experience was a major milestone in my internship. Standing by my initial L1 research steps and constructively advocating for the administrative access needed to solve the problem taught me how to navigate high-pressure situations in a professional environment. Also, discovering how memory exhaustion impactsWindows services deepened my knowledge of cloud resource allocation. Moving forward, I feel more confident utilizing monitoring tools like Azure Monitor while maintaining clear, professional communication with leadership.