ChatGPT for DevOps
Complex actions. Simple conversations
Kubiya.Ai
Error on Kubernetes namespace Observability:kubiya-ai-6c8dd8d967-7bhd8: 10.0.20.23
status CrashLoopBackOff
Restarts 6
Age 7m45sec
.png)
Dan (R&D)
@Kubiya What does it mean?
Kubiya.Ai
The CrashLoopBackOff error can occur due to various reasons — the wrong deployment of Kubernetes, liveness probe misconfiguration, and init-container misconfiguration. You can also bypass the error by creating a separate deployment with the help of a blocking command.
.png)
Sam (R&D)
@kubiya How many tickets are still not done in Jira for the current sprint?
Kubiya.Ai
With 3 days left into the current sprint 12 tickets are in-progress, 8 tickets are in QA and 2 tickets are in backlog
Sam (R&D)
@kubiya Which 2 tickets are in backlog?
Kubiya.Ai
The 3 tickets in backlog are:
FE-112 - Refactor page loader
OPS-89 - Setup a pipeline notification system
Sam (R&D)
@kubiya Move Jira ticket FE-112 to in-progress
Kubiya.Ai
Done. FE-112 was updated to in-progress
ChatGPT for DevOps
Need help? Tag Kubiya and speak your mind and Kubiya will do the rest.
It will understand the context of your request. What information is provided and what other questions to take to get full details. You can run multiple conversations in parallel and even switch context back and forth, just as you would with your DevOps team.
No one reads the docs?
With Kubiya they would
Knowledge taps into your organization’s knowledge systems and is able to automatically answer questions based on those docs as well as point to the docs myself.
Your docs are outdated? users can give you feedback using Kubiya and create knowledge entries (or tickets) based on human-to-human conversations.
Some of the knowledge systems support include
.png)
Joe (R&D)
@kubiya Summarize the conversation for me
Kubiya.Ai
The conversation was mainly about asking questions related to Kubernetes namespaces, pods, and deployments, as well as IAM users and MFA enablement. Joe was the one asking all the questions to a user named kubiya.
Joe asked kubiya about the available Kubernetes namespaces and running pods in certain namespaces. They also asked about the number of replicas available for deployment in a specific namespace and which IAM users did not have MFA enabled.
Lastly, Joe asked kubiya to summarize the conversation so far.
To explain the conversation in a way a 5-year-old would understand, Joe was asking questions about a special computer system called Kubernetes and how many things were running on it, how many copies of things there were, and who was allowed to use it. Then they asked the computer person to tell them a short story about everything they talked about.
Joe (R&D)
@kubiya How do I set up a VPN?
Kubiya.Ai
To setup a VPN you will need to make a request through the helpdesk portal.
Once approved you will get a username, a temporary password and a license to use in your VPN client. Here’s a short video and the link to the helpdesk site.
Reinforcement learning
with Human Feedback (RLHF)
Kubiya learns from user interactions to optimize the experience at the organization and at the user level. If a users chooses certain clusters or namespace most of the time, Kubiya will over time suggest them by default. if a 2nd or 3rd alternatives were chosen for a given natural language query, Kubiya will change the order of its suggestions.
RLHF is kept at the organization and the user level and is used to optimize the experience for other organizations (more about how we keep data secure here).
Finally the user can give feedback on responses to improve accuracy.
Access Control
Kubiya’s platform includes a built-in access control where you can define which users or groups get access to do specific actions. Users can request temporary or permanent access, through your chat tool, with Kubiya managing the TTL.
If you have policies defined in OKTA, LDAP, Kubiya supports Open Policy Agent (OPA).
Generative AI to Create Your
Custom Workflows
Sometime you want to create define the experience you provide end-users: you want proper guardrails on the information you present, filter options or ask questions in a way your users will understand.
Using generative AI you can describe the workflow you are interested in creating and Kubiya will generate a functional workflow, reducing the effort needed to just a few minutes.
Flexibility to create what you need
Some of the things you can do in Kubiya:
Ask Kubiya a question in another conversation
Run scheduled jobs
Create webhooks that start a conversation, and maybe pulls logs as well, on a given channel
Run long jobs asynchronously
Notify a group
Request another user’s approval
Trigger a specific action on a defined data and time
Analytics
Want to know which resources are used the most? Who makes most use of infrastructure or what documents are out-of-date?
Kubiya keeps track of everything and can give you a level of visibility you never had before.
Integrations
Kubiya has pre-built integrations and actions for many of the common engineering platforms. That allows you out-of-the-box as soon as setup is done to query many of your platforms and trigger actions. Have an in-house solution or don’t see an integration you need? Kubiya can integrate with any platform that has an API or SDK using a simple Python based action store. Most action stores takes minutes to create
How it works
Creating a ChatGPT-like for engineering platforms requires a complex setup. It involved authenticating users, local operators to perform actions, secure tunnels that communicate with Kubiya’s network, and a combination of large language models, each for a different purpose and sometimes multiple of these to optimize the results. Finally everything needs to be secure and performant.
