Linux Tools — kubectl Output Processing Reference

See also: k8s commands combined with linux.md for the full pipeline pattern with kubectl.

---

The Core 8 Tools

Every one of these shows up constantly when manipulating kubectl output. They're composable — chain them with |.

Tool	What it does	Example
awk '{print $N}'	Extract column N (whitespace-separated)	awk '{print $1}' = first column
sort	Sort ascending (alphabetical or by line)	sort = A-Z
sort -r	Sort descending	sort -r = Z-A
sort -n	Sort numerically (not alphabetically)	sort -n for numbers
tail -n +N	Skip first N-1 lines (start from line N)	tail -n +2 = skip header row
head -n N	Take first N lines	head -1 = top result only
grep "pattern"	Keep lines matching pattern	grep Running
grep -v "pattern"	Drop lines matching pattern	grep -v Completed
wc -l	Count lines	wc -l = number of pods
cut -d',' -f1	Cut by delimiter, take field N	cut -d',' -f2 = 2nd CSV field
uniq	Remove consecutive duplicate lines	Always sort first or it won't work
tee file.txt	Write to file AND pass through to stdout	Capture and continue pipeline

---

sort Variants

sort           # alphabetical ascending
sort -r        # alphabetical descending
sort -n        # numeric ascending (1, 2, 10 — not 1, 10, 2)
sort -rn       # numeric descending
sort -u        # sort + deduplicate (same as sort | uniq)
sort -k2       # sort by column 2 (whitespace-separated)
sort -k2 -n    # sort by column 2 numerically
sort -t',' -k2 # sort by column 2 using comma as delimiter

Why sort before uniq: uniq only removes consecutive duplicates. If duplicates are scattered through the input, uniq misses them. sort first groups all duplicates together so uniq can catch them.

---

tail -n +N — The Header Skipper

tail -n +2    # start from line 2 → skips the header row
tail -n +3    # start from line 3 → skips 2 rows

The +N syntax means "start from line N". Standard tail -N means "last N lines". These are opposite behaviours — + changes the meaning.

Most common use:

kubectl get pods -o wide | tail -n +2 | awk '{print $6}'
#                          ↑ skip column header row

---

awk — Column Extraction

awk '{print $1}'        # first column
awk '{print $6}'        # 6th column (IP in kubectl get pods -o wide)
awk '{print $1,$6}'     # columns 1 and 6, space-separated
awk '{print $1","$6}'   # columns 1 and 6, comma-separated
awk '{print $1"/"$2}'   # namespace/podname format

Why awk over cut: kubectl output uses variable-width spaces between columns. cut -d' ' breaks with multiple spaces. awk splits on any whitespace — handles kubectl's format correctly.

---

tee — Write to File AND Continue

kubectl get pods | tee pods.txt | grep Running
#                  ↑ writes all pods to file, passes ALL through to grep

vs > which terminates the pipeline:

kubectl get pods > pods.txt    # stops here — nothing passed to grep

Use tee when you want to save the raw output AND continue processing it. > when you just want to save.

---

Practical Patterns with kubectl

# Get pod IPs sorted, with header
echo "IP_ADDRESS" > pod_ips.txt
kubectl get pods -o wide | tail -n +2 | awk '{print $6}' | sort >> pod_ips.txt

# Count running pods
kubectl get pods | grep Running | wc -l

# Get pod names only (no headers)
kubectl get pods -o wide | tail -n +2 | awk '{print $1}'

# Pods on a specific node
kubectl get pods -o wide | grep node01 | awk '{print $1}'

# Top CPU pod - name and namespace
kubectl top pod -A --sort-by=cpu --no-headers | head -1 | awk '{print $1"/"$2}'

# Nodes sorted by memory usage (node_resource.md pattern)
kubectl top node --sort-by=memory --no-headers | head -1 | awk '{print $1}'

# All unique images running in a namespace
kubectl get pods -o wide | tail -n +2 | awk '{print $7}' | sort -u

---

Column Positions — Quick Reference

kubectl get pods -o wide:

$N	Content
$1	NAME
$2	READY
$3	STATUS
$4	RESTARTS
$5	AGE
$6	IP
$7	NODE

kubectl get nodes -o wide:

$N	Content
$1	NAME
$2	STATUS
$3	ROLES
$4	AGE
$5	VERSION
$6	INTERNAL-IP

kubectl top pod -A --no-headers (with -A namespace column shifts everything):

$N	Content
$1	NAMESPACE
$2	NAME
$3	CPU
$4	MEMORY

kubectl top pod --no-headers (no -A):

$N	Content
$1	NAME
$2	CPU
$3	MEMORY