andres

In addition the skills teach the agent about which kubectl commands are mutating so it doesn’t treat them and normal Bash() commands and asks you for permission first.

Hi @nodesocket! Thanks for your support! Your help ironing out the bugs after launch was incredibly helpful.

Thanks! I used Screen Studio for the demos (https://screen.studio). I had to pay for it but I couldn't find the zoom+highlight feature anywhere else.

Regarding Heroku - I loved Heroku the first time I used it so I set up my Kubernetes cluster to work in a similar way (`git push` to deploy to production). I know K8s has a reputation for being useful only for large deployments/large teams but in my experience this isn't true. You can run a single node cluster and use it to host multiple projects all sharing the same load balancer, or you can scale up to thousands of machines with many microservices working together internally. Would it be helpful for you to see a HOWTO for setting up a Heroku-like K8s cluster?

Yes, there's no index. It uses grep (powered by ripgrep) under the hood.

Kubetail is more lightweight than Loki/Grafana. It fetches logs using the Kubernetes API which means you can only see the current state of the cluster but you can use it without installing any additional software or provisioning storage. Our new search feature greps container log files on the nodes themselves as compared to Loki/Grafana that builds a full-text index in the background and queries that at search time (I think).

Stern is a CLI tool which more closely compares to the Kubetail CLI tool (as opposed to the Kubetail web interface). Currently, there's a lot of overlap between the two tools but Kubetail gives you more control over source filters and time. For example with the Kubetail CLI tool you can do queries like this:

    kubetail logs deployments/web \
      --since 2025-05-01T00:00:00Z \
      --until 2025-05-02T00:00:00Z \
      --zone us-east-1a \
      --with-ts \
      --with-node \
      --with-pod \
      --tail=100

Thanks for posting a link to Johan's project. I'm in touch with him to figure out the best way to avoid confusing users.

Yep! You can use Kubetail on your desktop (using the CLI tool) or you can install it directly in your cluster using helm:

    helm repo add kubetail https://kubetail-org.github.io/helm-charts/
    helm install kubetail kubetail/kubetail --namespace kubetail-system --create-namespace

    kubectl port-forward -n kubetail-system svc/kubetail-dashboard 8080:8080

Thanks! Kubetail doesn't cache or store logs itself. By default, it uses the Kubernetes API to fetch logs from your cluster and send them directly to your client (browser or terminal). If the "Kubetail Cluster API" is installed then it uses Kubetail's custom agent to do the same.

Thanks! Part of what enables us to make a helpful lightweight solution is that we're leveraging the Kubernetes API to give logs context without requiring extra configuration. It'd be great to generalize Kubetail but a lot of that depends on how cloud platforms evolve over the next few years. Do you use Kubernetes?

Thanks! Your comment made my day! It sounds like your use-case is similar to mine when I started working on the project. Now we have a community of contributors working on Kubetail so if you have time, stop by our Discord and let us know what else we can do to help (https://discord.gg/CmsmWAVkvX).

If you get a chance to try out the Kubetail CLI, I'd love to hear your thoughts! There's a lot of overlap with stern at the moment but we're planning on adding new features soon that will be unique (e.g. remote grep, system logs).

If you get a chance, please check out kubetail (https://github.com/kubetail-org/kubetail). It's an open source log viewer for Kubernetes. Currently you can use it to look at pod logs from beginning to end, grouped together by workload (e.g. Deployment, CronJob) with basic filtering available (e.g. node-id, AZ). It doesn't let you look at historical logs yet but that's where we're headed. We just launched so we're eager for feedback and we like to build out new features quickly.

Update: fixed in v0.1.1 https://www.npmjs.com/package/fancy-ansi

Created issue: https://github.com/kubetail-org/fancy-ansi/issues/2

Blink support would use CSS animations which requires a deeper level of integration than the other SGR parameters so I wanted to hold off until I could iterate over a solution with a user. If you'd like to use Fancy-ANSI with blink support, please let me know on GitHub and I can implement it fairly quickly.

Awesome! Great to hear it's useful for your project. Let me know if you have any questions or run into any issues.

Noted, thanks https://github.com/kubetail-org/fancy-ansi/issues/1

The library is focused on text styling so it only implements ANSI SGR parameters. I used this as a reference: https://en.wikipedia.org/wiki/ANSI_escape_code

The full list of supported SGR parameters is in the README: https://github.com/kubetail-org/fancy-ansi?tab=readme-ov-fil...

Let me know if you have any suggestions on how to make it clearer which subset of ANSI is supported.

I added "Fancy-ANSI" as ASCII art to the preview site: https://kubetail-org.github.io/fancy-ansi/

Great! Do you mean a screenshot? As far as I know, besides images, you can’t embed html in GitHub readmes.

Here's the github issue if you want to get notified when it gets fixed: https://github.com/kubetail-org/kubetail/issues/2

Awesome! Let me know if you run into any issues setting it up (andres@kubetail.com)

Thanks, just added one https://github.com/kubetail-org/kubetail

Awesome! I'm excited to hear how it goes. If you have any feedback please send me an email (andres@kubetail.com).

Currently, there's no persistence. I'll think about how to enable client-side.

Looks interesting! It's like tail -f on steroids.

Sorry about that! It's fixed now. Currently, kubetail displays the entire log history by default which can cause the frontend to hang if there's a lot of data. In the case of the kubetail-demo pod, there were a lot of messages from a user continuously retrying to make a websocket connection and the quantity of those messages crashed your tab. I've disabled logging for the app instance so viewing those pod logs won't crash your browser again. Upgrading the kubetail frontend so it can handle more data is next up on the to-do list.

Can you give me more details? Do you want to run kubetail locally but pull logs from your cluster remotely? Yes, this is possible.

That project is a bash script to tail Kubernetes logs from multiple pods at the same time. The name collision is a bummer. I found out about it a while after I bought the domain names (kubetail.com, kubetail.dev, etc.), got the social handles and invested a lot of time into coding and branding. If this project is successful and the naming is confusing for users we'll figure out a solution, if not then it's moot.

Thanks! Kubetail is super lightweight and requires zero maintenance so hopefully you can use it for some of your real-time logging needs. Let me know if you have any feature requests.

Kubetail runs in your cluster and uses the cluster’s own Kubernetes API to retrieve logs so the data never leaves your possession. Yes! Kubetail will always be open source. I think it’s important for users to know exactly what code is running inside their clusters especially if they’re granting it access to their private data.

Yes