From Devops to NoOps : can AI automate everything?
Video size:
Abstract
AI is revolutionizing CloudOps pushing us towards NoOps where infrastructure, deployments, security are self-managed with minimal human intervention. Discover how AI-driven automation boosts efficiency, reduces downtime, reshapes the role of DevOps engineers, while raising key security challenges.
Transcript
This transcript was autogenerated. To make changes, submit a PR.
Hey everyone.
My name Isman Carla.
I'm a senior technical lead and a cloud DevOps engineer with
seven years of work experience.
Thank you all for joining my talk.
Today we'll explore how AI is transforming how we build and deploy
the software from the collaboration world of DevOps to the future of no ops.
Initially we'll talk about how DevOps is evolved and how
we are transforming to NoOps.
Initially we started the DevOps to remove the gaps between development
and operation team so that we can accelerate the deployments from lower
environments to higher environments.
There are automation tools that improve efficiency, but human oversight remains
crucial in each step from continuous integration through continuous deployment.
However with the rise of no ops we are fully envisioning automation
with the minimum human interactions.
The primary reason behind this transformation is
because of a and MI models.
R. Driving this huge transformation in the world of DevOps.
Now we'll talk about how AI is playing a crucial role in DevOps
automation, AI powered continuous integration and continuous deployment.
Pipelines are built so that we can deploy the application with a
minimal human input and interaction.
This also helps us to find anomaly detection and also root cause analysis
using machine learning algorithms.
And also it offers a self-healing system that detects and
fixes issues automatically.
Also, a ops, our real time operational insights and alerts that helps us to
give more information about an issue that is happening in various environments.
This helps us to resolve the issues quickly and also to get the
application up with a minimal time.
And also a driven infrastructure management a has becoming smarter in
terms of provisioning infrastructure, for example if an infrastructure requires
an order, scaling it predictive analysis helps us whether to up the application or
down the application based on the traffic.
And also AI is helping us to manage Kubernetes clusters
with a minimal human inputs.
This helps us to minimize the cost and also improve the performance.
In the long run, AI is helping.
In terms of infrastructure as a code with intelligent valuation, this says that
while improving the code AI is helping us so that we can focus more on the
problem rather than syntax, for example, Datadog, which helps us to visualize.
The information from the clusters and also it also helps us to
understand how the infrastructure is behaving and also application
metrics, infrastructure metrics, and also the usage of the resources.
Now when coming to challenges and risks of New Ops, there are quite
number of challenges while implementing the new ops because A requires a
high quality data and data tuning.
If we rely completely on automation that might miss root causes for many problems.
And also security concerns are the biggest.
Discussion because developers cannot have the expertise in security in
finding anomalies and fixing them.
Even the automation tools have the vulnerabilities so that it might cause
security threat to, or it operations.
This is not just about shifting to tools.
It's just more about software development lifecycle.
Also the future of AI in DevOps.
This has become really efficient in generating a code and
also configuration changes.
For example, GitHub Coate helps us to generate Kubernetes configuration
files based on our prompt eight.
Reach all the code in our repos and generate a particular required AML file,
and then it deploys into the production.
Also hybrid model helps us combining human oversight with AI automation in
the future DevOps engineers transitioning into mostly a supervisors and strategies.
This is a visual representation of a driven evolution in cloud operations.
If you see we initially started with manual interventions with scripted
automations and DevOps collaborations.
With the time we moved to a driven continuous integration
and continuous deployment anomaly detection and also health self.
Healing systems also in smart infrastructure based on the predict
analysis, autoscaling can be take place based on traffic that
improvises cost as well as efficiency.
In the second and challenges and the ethics is more about data quality, AI
transparency, role and transformation now the future and no ops, which
we'll discuss in the next slide.
So now real question is that can AI truly replace human operations?
The answer is not yet.
The rise of NoOps doesn't end the signal of DevOps because DevOps can
be used for a more strategic and complex problems, whereas NoOps can be
used for a simple, automating tasks, full ops is still aspirational, but
human oversights remain critical.
DevOps is still evolving, but not disappearing.
However, with AI as a co-pilot, we can generate more sustainable and stable
application across all environments.
This is all about DevOps to NoOps can a automate everything.
Thank you all for joining my talk and I hope you all enjoy con 42 Site
Reliability Engineering Conference.
Thank you.