Supercharge Your DevOps Workflow with ChatGPT: Best Practices and Use Cases

Johannes is an DevOps engineer with passion for solving problem and learning new technologies. Johannes has the skill and experience about improving performance and also reducing infrastructure cost to help you achieve your goals.

Looking for Elite Digital Talent?
Hire vetted candidates within 2 weeks and save hiring cost up to 70%

Incorporating artificial intelligence into your DevOps workflow can be a game changer. ChatGPT is perfect for generating small scripts and answering basic questions, while the Playground tool is great for generating well-rounded configurations.

In this article, we will explore how ChatGPT, a language model based on the GPT-4 architecture, can be used to support and enhance DevOps processes, including best practices, use cases, and warnings.

Best Practices for Using ChatGPT in DevOps

ChatGPT is a powerful tool, but it's essential to use it correctly to get the best results.

Here are some best practices for using ChatGPT:

Carefully consider your prompts

The prompt you input is crucial to getting accurate and relevant results. Be as specific and concise as possible, and provide examples to help ChatGPT understand the context.

Have a conversation with ChatGPT

ChatGPT is designed to learn from interactions, so having a conversation with it can improve the accuracy and depth of its answers. Try asking follow-up questions or clarifying your prompts to get more detailed responses.

Verify the output

Always verify the output generated by ChatGPT to ensure its accuracy and depth. It's also a good idea to use multiple sources to cross-check the information.

Provide examples

Providing examples in your prompts can help ChatGPT generate more specific and relevant responses. For example, if you're asking for a script, provide an example of what the script should do.

Use ChatGPT for small scripts and configurations

ChatGPT is particularly useful for generating small scripts and configurations. It's not designed to replace DevOps engineers but can increase productivity and help unblock them in their work.

chatgpt for devops

Now, let’s move to essential use cases of ChatGPT for DevOps.

1) Generate Small Scripts

ChatGPT is an excellent tool for generating small scripts for various tasks. Here are some examples:

  1. Bash script: Write a Bash script that lists all files in a directory and subdirectories, sorted by modification date.
  2. Terraform script: Generate a Terraform script to deploy a Kubernetes cluster on AWS with a load balancer and a custom domain.
  3. Python script: Write a Python script for SSH onto a VM and check the memory usage then output the results to a file. The script should be scalable to work against multiple VMs.
  4. PowerShell script: Generate a PowerShell script that automates the installation and configuration of IIS on a Windows server.
  5. Log file script: Write a script that extracts information from a log file and creates a report with a summary of errors and warnings.

2) Generate Configurations

The OpenAI Platform Playground tool helps generate well-rounded configurations, and construct or correct any cloud infrastructure diagram and terraform code. Here are some examples:

  1. Apache Airflow deployment: Deploy Apache Airflow via Terraform using Helm.
  2. Database creation: Build your database from scratch.
  3. Event-driven application: Implement an AWS event-driven serverless story generation application.
  4. Cloud architecture: Create a functional cloud architecture with AWS and Azure resources in a specific region.

3) Troubleshooting with ChatGPT

ChatGPT can assist with troubleshooting, but it may not always provide new information. However, it can help save time and provide new perspectives. Here are some examples:

  1. Multiple clusters: Can I use one Thanos Receive endpoint to receive metrics from multiple clusters?
  2. Error resolution: How do I resolve this error in the [Component]: [error message]?
  3. Virtual machine sizing: Run commands inside the virtual machine to check the virtual machine sizing, as well as the network connectivity.

4) Automation with ChatGPT

ChatGPT can help automate tasks and improve efficiency in the DevOps workflow. Here are some examples:

  1. Create a BASH shell script that can restart a service hosted on multiple virtual machines (VMs).
  2. To resolve an error in the [Component] that displays an error message, run commands within the virtual machine to check its sizing and network connectivity.
  3. Write a single-line command/query to list all GCP instances that start with 'prod' and end with the keyword 'service'.
  4. Use an OpenShift command to patch the HPA of a particular namespace.
  5. Develop a Jenkins declarative pipeline code to clean up the app workspace, checkout code from GitHub, and build it using Maven.

5) Discover New Opportunities

It can help brainstorm ways to tackle business-impacted strategies such as cloud cost and security.

  • How can virtual machines be run like Pods and managed with Kubernetes?
  • Can you recommend a tool for backing up a Kubernetes cluster?
  • Let's brainstorm ways to address business concerns such as cloud costs and security. For example, "What are five ways to reduce AWS billing costs from $7000+ to under $2000?"

Or it can provide insights into what competitors are saying about a particular topic, such as:

  • Let's analyze what competitors are saying. For instance, "Why is Amazon being accused of being anti-competitive, particularly regarding cross-AZ data transfer fees?"

Service restart is a crucial task in DevOps. With ChatGPT, this task can be automated, making it easier and more efficient. One way to do this is by writing an automation process that can restart a service hosted on multiple virtual machines (VMs) using a BASH shell script.

The first step is to write the BASH shell script. This script should be designed to restart the service automatically. The script should be tested and debugged to ensure that it works as expected. Once the script is ready, it can be deployed on multiple VMs.

Next, you need to create a ChatGPT workflow to execute the BASH shell script on each VM. ChatGPT can help you create a workflow that will execute the script on each VM, one after the other. The workflow can be configured to handle errors and report progress.

To create the ChatGPT workflow, you need to define the inputs and outputs. The input is the list of VMs that need to be restarted. The output is the status of the script execution on each VM.

Here's what the ChatGPT workflow would look like:

  1. Retrieve the list of VMs from the input
  2. For each VM in the list:a. Execute the BASH shell script on the VMb. Capture the status of the script execution
  3. Return the status of the script execution for each VM as the output

Once the workflow is created, you can test it on a small set of VMs before deploying it on the entire fleet.


ChatGPT can help automate the deployment process by creating a workflow that can deploy code to multiple environments, such as development, testing, and production.

To create the ChatGPT workflow for deployment, you need to define the inputs and outputs. The inputs are the code to be deployed and the list of environments. The output is the status of the deployment for each environment.

Here's what the ChatGPT workflow would look like:

  1. Retrieve the code and list of environments from the input
  2. For each environment in the list:a. Deploy the code to the environmentb. Capture the status of the deployment
  3. Return the status of the deployment for each environment as the output

Once the workflow is created, you can test it on a small set of environments before deploying it on the entire fleet.

One crucial thing to remember is that ChatGPT is not a threat to DevOps engineers but rather a helpful tool to increase productivity and efficiency. While it may provide some level of automation, it cannot replace the skills and expertise of a human engineer.

Warnings about Using ChatGPT in DevOps

While ChatGPT can be a powerful tool for DevOps teams, some warnings should be taken into account.

Lack of human-like expression and feeling.

The output generated by ChatGPT cannot communicate thoughts and emotions in the same way that humans can. This can make it difficult to create content that resonates with users on a deeper level.

Produce low-quality content that appears relevant and coherent but lacks accuracy or depth

This can be due to limitations in the algorithm or training data. Furthermore, the human review process for ChatGPT content can vary widely in quality, and reviewers may not have the necessary expertise to evaluate the output accurately. This can result in important errors or inaccuracies being missed.

Insufficiency of unique insights into a given topic.

While it can summarize a topic, it does not offer any subjective perceptions or personal experiences that humans can bring to the table. In some cases, ChatGPT may even hallucinate facts or fabricate answers to persuade reviewers that its content is correct.

The inappropriate result when a direct answer is needed

ChatGPT tends to approach topics from multiple angles. This can result in lengthy web pages that may not be suitable for site visitors seeking a straightforward answer.

Despite these drawbacks, ChatGPT is still in training and is expected to improve in the future. However, until its limitations are fully understood, it is important to keep humans in the loop and have them review ChatGPT outputs before they are used in critical applications. This is especially important for code generation or high-stakes domains, where the limitations of the system need to be taken into account.

Biases and errors in the training data

ChatGPT is trained on a massive corpus of text data, which may contain biases and errors. This can lead to the model producing inaccurate or biased outputs, especially when the training data is not representative of the target domain.

Difficulty in interpreting outputs

ChatGPT outputs can be difficult to interpret, especially when they are not consistent with the user's expectations or domain knowledge. This can lead to misinterpretation or misuse of the outputs, and potentially cause errors or security risks.

Dependency on cloud services

ChatGPT requires significant computing power and infrastructure to run, which may be expensive and require a dependency on cloud services. This can lead to additional costs and potential risks associated with cloud service providers.


  1. How accurate are ChatGPT outputs?Answer: The accuracy of ChatGPT outputs depends on several factors, such as the quality of the training data, the model's hyperparameters, and the specific task at hand. It is important to evaluate and monitor the outputs regularly to ensure that they are accurate and reliable.
  2. Can ChatGPT be used for sensitive or confidential information?Answer: ChatGPT outputs can potentially reveal sensitive or confidential information, so it is important to implement security and privacy measures, such as encrypting the input and output data and restricting access to the model.
  3. What are some potential biases in ChatGPT outputs?Answer: ChatGPT outputs can potentially contain biases, such as gender or racial biases if the training data contains these biases. It is important to carefully select and preprocess the training data to minimize these biases.
  4. Is ChatGPT suitable for all DevOps use cases?Answer: ChatGPT is suitable for some DevOps use cases, such as natural language interfaces and knowledge management, but may not be suitable for all use cases. It is important to carefully evaluate the suitability of the tool for each use case.
  5. Can ChatGPT be used without a cloud service?Answer: ChatGPT requires significant computing power and infrastructure to run, so it may be difficult to use without a cloud service. However, there are some options for running the tool on local hardware or with a smaller cloud service provider.


In conclusion, ChatGPT can be a valuable tool for DevOps engineers as it has the potential to enhance productivity and facilitate problem-solving. Nonetheless, it's crucial to acknowledge the limitations of ChatGPT and not rely on it for unsupervised DevOps implementation.

One of the main issues with ChatGPT is its inherent biases that can lead to inaccuracies and shortcomings. Additionally, it struggles to generate content that feels natural or provides unique insights.

To optimize the output of ChatGPT, users should provide detailed prompts to generate relevant and high-quality content that meets their expectations. Engaging a professional to write a prompt can significantly impact the quality of the follow-up answers provided by ChatGPT.

Therefore, it's essential to exercise caution while using ChatGPT and be aware of its limitations. Embracing technology and using it to your advantage can improve efficiency, but it's vital to balance the benefits with potential drawbacks.

If you're a company looking to hire a DevOps engineer to help with ChatGPT integration, it's crucial to find the right candidate for the job. That's where Talentport comes in - as the best SEA tech talent provider.

We can help you find the perfect DevOps engineer to increase your business ROI by up to 300%. Don't hesitate to contact us now for assistance in finding the right fit for your hiring needs.

Grow your business faster with
Top Tier Remote Talents

Get 300% ROI from your hiring budget