Forem: Aidas Petryla

We Almost Didn’t Launch — Until We Cut the Product in Half

Aidas Petryla — Mon, 20 Apr 2026 13:12:28 +0000

🚀 We almost didn’t launch.

Not because the product was impossible.Not because the team wasn’t capable.

Because the scope kept growing… while our capacity didn’t.

A year ago, launching soci.lt felt like chasing a horizon:– People had jobs, studies, life– Capacity was unpredictable– “Just one more feature” kept sneaking in– Timeline kept slipping

And the worst part?Until launch — nothing moves. No users, no feedback, no signal. Just… waiting.

So I did something uncomfortable.

I cut.

Hard.

Users could register… but not delete accounts.You could create data… but not delete it.Some profile fields? Not editable.Logout? Not critical.

If it didn’t block the core flow, it was gone.

The rule became simple:If this can be done the day after launch — it’s not a launch requirement.

That’s when things changed.

Progress became visible.The backlog started shrinking for real.Energy shifted.

But the real challenge wasn’t the backlog — it was discipline.

“Must-have” features kept coming (from everywhere, including my own brain).Each one had to go through the same filter:

Is this truly blocking us from going live?

If not → Day 2.

Another move that helped:We aligned development with business reality.

When it looked like we needed ~3 weeks to finish, and onboarding would also take ~3 weeks — we overlapped them.

Started onboarding before we were fully done.

Risky? Yes.Worth it? Absolutely.

We launched earlier than expected.

And here’s the funny part:

Some of those “critical before launch” features…are still not built today.

Turns out:Doing something manually 1–2 times a yearis cheaper than a week of engineering time.

A lot of projects don’t fail because they’re bad.

They fail because they never cross the line.

Like ships stuck in the harbor, endlessly polished, upgraded, prepared…but never allowed to sail.

What helped us actually launch:

Define a real feature list for launch
Next day — cut it again (ruthlessly)
Map dependencies, estimate, timeline (basics, but done properly)
Protect priorities when chaos hits (because it will)
Align with business timing — not just engineering readiness

And one more, which I learned the hard way:

Prefer reversible decisions early (features)Be careful with irreversible ones (architecture)

This approach is not universal.

Sometimes you should start fully manual — spreadsheets, forms, emails — and evolve from there.

But if you’re stuck in “almost ready” for months (or years)…you might not need more effort.

You might need less scope.

Harsh truth: Many products die in the shadows of perfection.

Better to launch something imperfect that breathesthan something perfect that never lived.

The Opposite of Loneliness

Aidas Petryla — Mon, 14 Jul 2025 09:41:10 +0000

After 30+ years of feeling lonely, I've been thinking about what's missing.
Yes, there are apps for meeting people—Meetup, Bumble BFF, and others. But they don't quite capture what I'm searching for: a space to find people who share our passions, our curiosities, a desire to build RC aircraft or to have that Harry Potter marathon we've been putting off.

I'm searching for (or dreaming of building) a platform that:

Puts community over profit — No subscription tiers hiding basic features. No algorithms designed to keep you scrolling instead of connecting. A place where you can message interesting people instantly without hitting 3-messages-per-day limits (you don't have limits saying "hello" to people on the street, right?). You can search for people by location, interests, etc. without limits!
Connects through shared purpose — Find people who want to collaborate on projects, learn together, or simply share genuine interests. Not just "swipe right if you like hiking."
Builds lasting relationships — Success means people forming real friendships, not deleting the app after a single match.
Respects authenticity — No pressure for perfect photos or curated personas.
Grows organically — Ideally open-source, built by the community it serves.

I believe loneliness is one of humanity's biggest challenges—one that money, fame, or even AI can't solve. We're social beings. We need each other.

If you relate, or know any tools that already do this well, I'd love to hear from you. If you're interested in building something like this together — I'm more than happy to grab a cup of tea and jump on a call to connect and discuss more. ^^

Concurrent and Parallel Programming in Python (course)

Aidas Petryla — Mon, 22 Jul 2024 07:02:00 +0000

Just finished an excellent course on Concurrent and Parallel Programming in Python, and I'm excited to share my experience!
As a lead engineer, I'm always looking for ways to optimize performance and expand my knowledge. This course by Max S on Udemy was a fantastic refresher on async programming, parallel computing, and threading.

What stood out:

Clear explanations of complex concepts
Practical, real-world examples
Hands-on coding opportunities

The instructor breaks down threading, multiprocessing, and asynchronous programming in Python with ease. We built a multi-threaded program that fetches data from the internet, parses it, and saves it to a local database - a common scenario many of us face in our daily work.

Key takeaways:

Optimizing IO-bound operations with multi-threading and async programming
Leveraging multiprocessing for CPU-bound tasks
Combining async and multiprocessing for maximum efficiency

Whether you're looking to speed up data processing, improve API performance, or just refresh your understanding of concurrent programming in Python, I highly recommend this course.

Have you taken any courses lately that improved your coding skills? Let's discuss in the comments!

Docker log rotation

Aidas Petryla — Mon, 15 Jul 2024 14:25:46 +0000

Today, I deleted 130 GB of logs. I don't know who started those Docker containers, but let me share a valuable concept: log rotation.

For example, one can define it in daemon.json:

{
  "log-driver": "json-file",
  "log-opts": {
    "max-size": "10m",
    "max-file": "3",
  }
}

or directly when starting a container:

docker run \
  --log-driver=json-file \
  --log-opt max-size=10m \
  --log-opt max-file=3 \
  my-image

or even as an Ansible script:

- name: Start Grafana loki
  docker_container:
    name: my-container
    image: my-image
    state: started
    log_driver: json-file
    log_options:
      max-size: "10m"
      max-file: "3"

There's no need to store 130 GB of logs. Really. :)

Discovering the Power of Atlassian Resources

Aidas Petryla — Sat, 29 Jun 2024 09:21:49 +0000

Recently, I explored the extensive resources provided by Atlassian on Agile, DevOps, and more. Their comprehensive material not only refreshed my understanding of Agile and Scrum methodologies learned at Red Hat, but also expanded my knowledge on critical topics like various Git workflows, DevOps culture, and the Scaled Agile Framework (SAFe).

Practical Tips from Atlassian

Consider Gitflow for Large, Complex Projects: Atlassian's guide on Git workflows offers valuable insights into enhancing your version control strategy. The Gitflow model, in particular, is useful for managing larger projects, ensuring a streamlined and efficient development process. It might become The Solution if your current workflow isn't scaling well.
DevOps Best Practices: Their insights into DevOps culture emphasize the importance of continuous integration and continuous deployment (CI/CD). One key takeaway for me was how crucial it is to address the cultural aspects of DevOps. Even the best processes won't yield results if the team culture isn't supportive.
Scaled Agile Framework (SAFe): For organizations looking to scale Agile practices, Atlassian’s resources on SAFe provide practical advice on aligning multiple teams towards common objectives. This can greatly improve overall productivity and cohesion. It's worth exploring if you're finding it challenging to maintain alignment as your team grows.

I'm interested in Your experience with Agile, DevOps and innovation in the organizations. Any stories to share or tips to suggest? :)

🎉 Just completed the "Kubernetes Fast Track" course on Udemy! 🚀

Aidas Petryla — Mon, 24 Jun 2024 09:28:29 +0000

This course has further solidified my Kubernetes skills and broadened my understanding of container orchestration. It was a great, easy to understand, and fast-paced course, which I'm recommending to people new to Kubernetes or those who haven't worked with it for a while and would like to refresh the basics.

Combining this with my extensive OCP experience and my current role managing multiple Nomad-Consul-Vault clusters, I'm more equipped than ever to tackle complex DevOps challenges. 💪

Kubernetes Fast Track

Understanding S3 Bucket Replication vs. CloudFront

Aidas Petryla — Sat, 15 Jun 2024 12:39:05 +0000

If you're diving into AWS, understanding when to use S3 Replication versus CloudFront is essential.

Amazon S3 Replication allows you to replicate objects across different regions, providing near-real-time sync. This is beneficial when you need:

Data redundancy: Ensuring high availability and disaster recovery.
Compliance requirements: Storing data in multiple geographical locations.
Low latency access: For applications where data access speed is critical across different regions.

AWS CloudFront is a global Content Delivery Network (CDN) designed for:

Fast content delivery: Reducing latency by caching content at edge locations.
Improving user experience: Quick load times for websites, videos, and APIs.
Enhanced security: Features like DDoS protection.

When to Use S3 Replication:

You need reliable, scalable storage for replicated data.
Data redundancy and compliance are priorities.
You require low-latency data access in specific regions.

When to Use CloudFront:

Speed and performance for delivering content globally.
Improving user experience with low latency.
Security and performance enhancements for web applications.

Combining S3 and CloudFront can leverage the strengths of each: use S3 for storage and replication, and CloudFront for efficient, global content delivery.

Making Meetings Matter

Aidas Petryla — Mon, 10 Jun 2024 04:29:00 +0000

In today's fast-paced work environment, productive meetings are essential to keep everyone aligned and moving forward. Yet, too often, meetings can become unproductive, leaving participants feeling unheard and unclear about the next steps. Here are some tips to ensure your meetings are efficient and effective:

Set Clear Objectives: Before the meeting, define what you aim to achieve. Share the agenda with participants in advance to ensure everyone comes prepared.
Active Listening: It's not enough to just speak; active listening is crucial. Make sure everyone feels heard by rephrasing and summarizing key points during discussions. This helps confirm understanding and align perspectives.
Encourage Participation: Foster an inclusive environment where everyone feels comfortable contributing. This not only enriches the discussion but also ensures that diverse viewpoints are considered.
Utilize Templates: Tools like Confluence offer a variety of templates designed to streamline meeting processes. These templates can help structure your meetings, keeping them focused and productive.
Action Items and Follow-Up: Conclude meetings with clear action items and assign responsibilities. Ensure there is a follow-up plan to track progress and address any issues promptly.
Time Management: Respect everyone's time by sticking to the agenda and managing the meeting duration effectively. If necessary, schedule follow-up meetings rather than extending beyond the allocated time.

By implementing these strategies, you can transform your meetings into productive sessions that drive results. Effective communication and structured processes are key to ensuring that everyone remains on the same page, both during and after the meeting.

What are your tips for holding productive meetings? Share your thoughts and let's learn from each other!

Unlocking the Power of AWS Auto Scaling: A Game-Changer for DevOps

Aidas Petryla — Sat, 01 Jun 2024 09:41:21 +0000

In the dynamic world of cloud computing, staying ahead of the curve is crucial. As I delve deeper into AWS, I've discovered a transformative feature that significantly enhances how we manage our infrastructure: AWS Auto Scaling.

What is AWS Auto Scaling?

AWS Auto Scaling is a powerful service that allows you to automatically adjust your application's capacity to maintain steady, predictable performance at the lowest possible cost. It provides a hands-off approach to scaling your instances, eliminating the need for manual intervention and constant monitoring.

Key Capabilities of AWS Auto Scaling

Automatic Scaling

AWS Auto Scaling automatically adjusts the number of EC2 instances in response to changes in demand. Whether you're experiencing a sudden spike in traffic or a lull, Auto Scaling ensures your application always has the right amount of resources.

Dynamic Scaling Policies

With AWS, you can define dynamic scaling policies based on various metrics like CPU utilization, memory usage, or custom CloudWatch metrics. These policies allow your application to respond to real-time changes efficiently, scaling up during peak times and scaling down during off-peak periods.

Scheduled Scaling

For predictable traffic patterns, AWS allows you to schedule scaling actions. This means you can pre-define scaling activities to align with anticipated load changes, ensuring optimal performance without the need for manual adjustments.

Predictive Scaling

AWS Auto Scaling uses machine learning to predict future traffic patterns and proactively adjust your resources. This cutting-edge feature minimizes response times and maximizes cost efficiency by forecasting demand and preparing your infrastructure accordingly.

Integrated Monitoring and Alarms

AWS integrates with CloudWatch to provide comprehensive monitoring and alerting capabilities. You can set up alarms to notify you of any unusual activity, ensuring you're always in control and informed about your application's performance.

Why AWS Auto Scaling is a Game-Changer

Cost Efficiency: By scaling down during low demand periods, you save on unnecessary infrastructure costs.
Enhanced Performance: Automatic scaling ensures your application remains responsive and performant, even under varying loads.
Reduced Operational Overhead: Eliminates the need for constant manual adjustments, freeing up your team to focus on more strategic tasks.
Scalability: Easily handle traffic spikes and growth without any manual intervention, providing a seamless experience for your users.

Real-World Impact

Imagine running an e-commerce site that experiences a surge in traffic during holiday sales. With AWS Auto Scaling, your infrastructure can automatically scale to handle the increased load, ensuring a smooth shopping experience for your customers. Post-sale, it scales down to save costs, maintaining efficiency.

Can’t Reach Your Manager? Here’s How to Still Do Your Best Work!

Aidas Petryla — Sun, 19 May 2024 07:37:01 +0000

Wouldn't it be great to reach out to your manager and get a response anytime you need it? Unfortunately, that isn't always the case. Many times, the manager is so busy that it's hard to get a response at all.

Here's what I do when it's hard to reach out:

Seek information from others or through different channels. Usually, there's more than one person who can help. Choosing the path of least resistance might be the solution.
Work independently whenever possible. Sometimes it's faster to complete the work yourself and then seek feedback rather than asking many less important questions, which can lead to significant delays.
When you must check in with your manager but they often overlook your messages, try this approach: "I intend to do <...>. If you'd like me to change my course of action, please inform me by EOD."
Highlight urgency when necessary. There are times when you can't proceed without your manager's input. In such cases, it’s crucial to let them know that you are blocked without their response. Sometimes a manager's interaction is essential, and that’s part of their job.
Discuss and agree on communication preferences. Transparent conversations about preferred communication methods, levels of reporting, and which decisions do or don't require approval can be beneficial. This ensures that both sides are on the same page.

What do you think? Are there any strategies here that you’ve used successfully, or any that you would avoid at all costs?

Unlocking AI Superpowers in DevOps: Master Prompt Engineering to Revolutionize Your Workflow!

Aidas Petryla — Sun, 28 Apr 2024 18:15:38 +0000

As a Lead DevOps Engineer who continuously explores ways to work smarter and enhance our operations with AI, I recently completed a course on Prompt Engineering that offered some noteworthy insights into leveraging ChatGPT effectively.

Key learnings:

AI as a Partner: The course emphasized the concept of AI as a partner, not a threat. By understanding how to communicate effectively with tools like ChatGPT, we can enhance our problem-solving capabilities.
Prompt Engineering Skills: Structural patterns in prompts can drastically change the output quality. As engineers, we appreciate systematic approaches, and this is no different when instructing an AI.
Efficiency in Automation: ChatGPT can be utilized to generate code snippets, automate repetitive communications, and even aid in complex troubleshooting scenarios, all by mastering the art of prompt crafting.
Mindset for Innovation: AI requires a different mindset. It's about guidance, iteration, and continuous learning—skills that are at the core of DevOps philosophy.

Tips for fellow DevOps professionals:

Be Specific: Tailor your prompts to be as detailed as possible. The more context you give, the better the AI can assist you.
Iterate and Learn: AI won't always get it right the first time. Use each interaction as a learning opportunity to refine your prompts.
Use AI for Scale: ChatGPT can help scale your operations by handling tasks like generating documentation, providing first-level support answers, and even creating templates for infrastructure as code.

By integrating these practices, we can make our workflows more efficient and stay ahead in the rapidly evolving field of AI. Let’s embrace the change and lead the transformation in our DevOps operations.

Would love to hear how others are integrating AI into their workflows! Feel free to share your experiences.

https://www.udemy.com/course/chatgpt-prompt-engineering/

ChatGPT for IT Workers (short Udemy course review)

Aidas Petryla — Thu, 18 Apr 2024 09:20:49 +0000

Just completed the "ChatGPT for IT Workers" course and I'd like to share my thoughts.

This course is particularly beneficial for individuals who are new to ChatGPT. It provides a wealth of examples demonstrating how this tool can be applied in the IT sector.

However, those who have been utilizing ChatGPT for a while might not find it as valuable. The course focuses on introductory content and does not explore more advanced concepts or complex prompts.

https://udemy.com/course/chatgpt-for-it-workers/