# Packages
# README
k tool
kubectl wrapper and utilities
Basic usage
Type k instead of kubectl and save 6 keystrokes.
The command to be run will be printed to stderr.
$ k get nodes
kubectl get nodes
NAME STATUS ROLES AGE VERSION
ip-10-10-1-204.us-west-2.compute.internal Ready <none> 60m v1.18.14
Avoid typing -n <ns> over and over by exporting NS
$ export NS=some-long-ns-name
$ k get pods
kubectl -n some-long-ns-name get pods
NAME READY STATUS RESTARTS AGE
some-pod 1/1 Running 0 10m
Utilities
mux
Multiplex exec command output across Pods onto stdout
Usage: k mux exec <command>
Useful for gathering output from Pods by selector. NS is still set from above.
$ SELECTOR=app=envoy k mux exec echo hi
kubectl -n some-long-ns-name exec envoy-65f74d8949-5sk9b echo hi
hi
kubectl -n some-long-ns-name exec envoy-65f74d8949-5sk9b echo hi
hi
kubectl -n some-long-ns-name exec envoy-65f74d8949-5sk9b echo hi
hi
The command is sent to stderr so output can captured in the clipboard or a file
$ SELECTOR=app=envoy k mux exec echo hi | pbcopy
kubectl -n some-long-ns-name exec envoy-65f74d8949-5sk9b echo hi
kubectl -n some-long-ns-name exec envoy-65f74d8949-5sk9b echo hi
kubectl -n some-long-ns-name exec envoy-65f74d8949-5sk9b echo hi
$ pbpaste
hi
hi
hi
rex
Manipulate stdin with a regular expression and output capture groups
Usage: k rex <regular expression> <capture group 1> [capture group 2] [capture group 3]...
$ echo 'abc 123\ndef 456'
abc 123
def 456
$ echo 'abc 123\ndef 456' | k rex '\w+' 0
abc
def
$ echo 'abc 123\ndef 456' | k rex '\d+' 0
123
456
$ echo 'abc 123\ndef 456' | k rex '(\w+) (\d+)' 2 1
123 abc
456 def
Top pods example
$ k top pods
kubectl -n some-long-ns-name top pods
NAME CPU(cores) MEMORY(bytes)
envoy-65f74d8949-5sk9b 33m 75Mi
envoy-65f74d8949-87h54 38m 77Mi
envoy-65f74d8949-rrx9z 36m 75Mi
$ k top pods | grep envoy | k rex '((\w|-)+) +(\d+)' 3 1 | sort -rn
kubectl -n some-long-ns-name top pods
38 envoy-65f74d8949-87h54
36 envoy-65f74d8949-rrx9z
33 envoy-65f74d8949-5sk9b
numstats
Run basic statistics on stdin
Usage: k numstats (assumes numerical input on stdin)
$ echo '1\n2\n3'
1
2
3
$ echo '1\n2\n3' | k numstats
count: 3
sum: 6
avg: 2
𝜎: 1
max: 3
min: 1
Look for even load distribution on Pods
$ k top pods | grep envoy | k rex '(\w|-)+ +(\d+)' 2 | k numstats
kubectl -n some-long-ns-name top pods
count: 151
sum: 95139
avg: 630
𝜎: 207
max: 1131
p99: 1130
p95: 963
p50: 626
min: 523
uniq
Print unique values from stdin
$ echo 'one\ntwo\ntwo\nthree\nthree\nthree'
one
two
two
three
three
three
$ echo 'one\ntwo\ntwo\nthree\nthree\nthree' | k uniq
one
two
three
count
Similar to uniq but count stdin and sort in descending order
Usage: k count
$ echo 'one\ntwo\ntwo\nthree\nthree\nthree'
one
two
two
three
three
three
$ echo 'one\ntwo\ntwo\nthree\nthree\nthree' | k count
3 three
2 two
1 one
set
Set operations (intersection, difference, union) on 2 files
Usage: k set <int|diff|union> <file 1> <file 2>
Given file 1 at ~/set1 with
a
b
c
and file 2 at ~/set2 with
c
d
e
$ k set int ~/set1 ~/set2
c
$ k set diff ~/set1 ~/set2
a
b
$ k set union ~/set1 ~/set2
a
b
c
d
e
time
Time deltas from now or a RFC 3339 provided time
No arguments prints usage and the current local time which is convenient to then copy and alter as the second argument (see below)
$ k time
USAGE: time <duration string> [2022-06-02T15:04:37-06:00]
Replace the -06:00 with Z to get UTC
If the current date and time was 2022-01-01T00:00:00Z then this would add an
hour to it
$ k time 1h
2022-01-01T01:00:00Z
If the current data and time is no longer 2022-01-01T00:00:00Z but you want to
calculate time deltas from then you can pass it as the second argument. Negative
durations can also be used
$ k time -1h 2022-01-01T00:00:00Z
2021-12-31T23:00:00Z
The available units for the duration are described in time.ParseDuration
Notably d for day is missing which is often inconvenient
Extensibility
Drop new command files into cmd/ and register them like this
func init() {
Register("mycmd", mycommand)
}
func mycommand() {
}
Or import "github.com/phylake/k/cmd" and call cmd.Register()