Skip to content

Latest commit

 

History

History
65 lines (48 loc) · 2.1 KB

README.md

File metadata and controls

65 lines (48 loc) · 2.1 KB
Raw

ktop (Kubernetes Top)

Top for Kubernetes container metrics in mostly realtime. ktop will connect to the Kubernetes cluster set in your Kubernetes config KUBECONFIG, it connect to the Kubernetes metrics client and polls for new changes every few seconds.

ktop starts an interactive terminal that will show the current container CPU and Memory metrics. You can order by CPU or Memory usage and filter based on the namespace, pod or container name.

Once a row is hightlighted, you will be able to see the Kubernetes resource requests and limits down the bottom, this will show what requests are currently set for that container.

ktop allows you to take a snapshot of the current metrics and compare all new metric results to that snapshot data set, you should see the CPU and MEM columns change colour depending on whether the new data is higher or lower than the snapshot data. It will show you the snapshot result vs the latest updated result.

Installing

Install the usual Go way:

$ go get -u github.com/vishen/ktop

Running

$ ktop

Bindings

Key Binding

  • Any ascii character entered will be used to filter
  • 1 - Order by CPU usuage descending
  • 2 - Order by CPU usage ascending
  • 3 - Order by Memory usage descending
  • 4 - Order by Memory usage ascending
  • UP - move up the list
  • DOWN - move up the list
  • SPACE - Snapshot of the current data to compare all new data with
  • ESC - Quits the application

Mouse Binding

  • Left click - follow that particular container
  • Right click - stop following the container

TODO

- Remove all magic +1, +5 numbers
- Fix hack for resizing; currently we just reduce pod and container name to reasonable size
- Remove locks from display
- Scrolling for when lines is bigger than terminal
    - Not always a problem as you can search for containers
- Configurable kubeconfig in cmdargs
- Change watch time will in interactive mode
- Show current cluster name in application heading
- Add the same for nodes?
+ highlight any recent changes
    - show only the delta change, currently shows the whole number
- Add new page to expand on pod / node info