Pods

Monitor and manage Kubernetes pods - view logs, check health, restart containers, and troubleshoot issues

Overview

The Pods page provides detailed monitoring and management of Kubernetes pods. View real-time container logs, check health probes, track restart history, and troubleshoot pod issues.

List View

Features

Live Status

Real-time pod status with phase indicators

Container Logs

View logs for any container directly from the list

Health Monitoring

Track liveness and readiness probe status

Restart Tracking

Monitor container restart counts and history

Pod Status

Pods can be in one of several phases:

Running

Pod is running normally. All containers started successfully.

Indicator: Green badge with “Running”

What it means: The pod has been bound to a node and all containers have been created. At least one container is running or is in the process of starting or restarting.

Pending

Pod is waiting to be scheduled or containers are being created.

Indicator: Yellow badge with “Pending”

Common reasons:

• Waiting for node assignment
• Pulling container images
• Insufficient cluster resources
• Volume mounting in progress

Failed

Pod has failed and will not restart.

Indicator: Red badge with “Failed”

Common reasons:

• Container exited with error code
• Image pull failed permanently
• Node failure
• Resource constraints

Succeeded

Pod completed successfully (typically from Jobs).

Indicator: Blue badge with “Succeeded”

What it means: All containers in the pod have terminated successfully and will not be restarted.

CrashLoopBackOff

Container is repeatedly crashing.

Indicator: Red badge with “CrashLoopBackOff”

Common reasons:

• Application startup failure
• Missing dependencies
• Configuration errors
• Resource limits too low

ImagePullBackOff

Container image cannot be pulled.

Indicator: Orange badge with “ImagePullBackOff”

Common reasons:

• Image doesn’t exist
• Registry authentication failed
• Network connectivity issues
• Wrong image name/tag

Evicted

Pod was evicted from the node.

Indicator: Red badge with “Evicted”

Common reasons:

• Node out of resources
• Exceeded pod memory limit
• Node maintenance
• Priority preemption

Table Columns

Column	Description
Name	Pod name (clickable to view details)
Status	Current pod phase badge
Node	Node where pod is running
IP	Pod IP address
Restarts	Total container restart count
Age	Time since pod creation
Actions	Quick action buttons (View Logs, Restart)

Status Filters

Filter pods by status:

All

Show all pods in the namespace

Running

Show only healthy running pods

Pending

Show pods waiting to start

Failed

Show failed pods

CrashLoopBackOff

Show pods repeatedly crashing

Detail View

Click any pod name to view comprehensive details.

Overview Section

1. Basic Information

   • Namespace: Current namespace
   • Node: Node assignment with link to node details
   • IP Address: Pod IP and Host IP
   • Service Account: ServiceAccount used by pod
   • QoS Class: Quality of Service class (Guaranteed, Burstable, BestEffort)
2. Status Information

   • Phase: Current pod phase
   • Conditions: Pod conditions (Initialized, Ready, ContainersReady, PodScheduled)
   • Start Time: When pod started running
   • Labels: Key-value labels
   • Annotations: Pod annotations
3. Resource Usage

   • CPU Requests/Limits: Total for all containers
   • Memory Requests/Limits: Total for all containers
   • Storage: Persistent volume claims

Containers Section

View details for each container in the pod:

Container Information:

• Name: Container name
• Image: Full container image path
• Image Pull Policy: Always, IfNotPresent, Never
• State: Running, Waiting, Terminated
• Ready: Whether container passed readiness checks
• Restart Count: Number of times container restarted
• Ports: Container ports and protocols

Click the “View Logs” button next to any container to open the logs viewer

Container Logs

Real-time log viewer with advanced features:

Features:

• Container Selection: Switch between containers in multi-container pods
• Line Limit: Choose number of log lines (100, 500, 1000, 5000)
• Auto-refresh: Logs update automatically
• Syntax Highlighting: Colored output for errors, warnings, JSON
• Search: Filter log lines by keyword
• Download: Export logs to file
• Timestamps: Toggle timestamp display

Example Usage:

# View last 100 lines
[Container: app] [Lines: 100]

2024-11-28T10:15:23.456Z INFO  Server started on port 3000
2024-11-28T10:15:24.789Z INFO  Connected to database
2024-11-28T10:15:25.123Z WARN  Cache miss for key: user:123
2024-11-28T10:15:26.456Z ERROR Failed to connect to Redis

Logs are fetched from the Kubernetes API and may have delays depending on cluster configuration

Health Checks

View liveness and readiness probe configuration and status:

Liveness Probe

Determines if container is alive and should be restarted:

HTTP GET

Performs HTTP request to specified path and port

livenessProbe:
  httpGet:
    path: /healthz
    port: 8080
  initialDelaySeconds: 30
  periodSeconds: 10

Exec Command

Executes command inside container

livenessProbe:
  exec:
    command:
    - cat
    - /tmp/healthy
  initialDelaySeconds: 5
  periodSeconds: 5

TCP Socket

Attempts TCP connection to specified port

livenessProbe:
  tcpSocket:
    port: 8080
  initialDelaySeconds: 15
  periodSeconds: 20

Readiness Probe

Determines if container is ready to accept traffic:

• Success: Container added to service endpoints
• Failure: Container removed from service endpoints
• Configuration: Similar to liveness probe

Failed readiness probes don’t restart the container - they only remove it from service load balancing

Startup Probe

For slow-starting containers:

• Disables liveness/readiness checks until first success
• Prevents premature restart during startup
• Useful for legacy applications with long initialization

Container Restart History

Track when and why containers restarted:

Information Displayed:

• Timestamp: When restart occurred
• Reason: Why container restarted (Error, OOMKilled, etc.)
• Exit Code: Container exit code
• Message: Detailed error message
• Signal: Termination signal (if applicable)

Common Exit Codes:

Code	Meaning
0	Success (normal termination)
1	General error
125	Container failed to run
126	Command cannot be invoked
127	Command not found
137	SIGKILL (OOMKilled or killed externally)
139	SIGSEGV (segmentation fault)
143	SIGTERM (graceful shutdown)

Events

Recent events related to this pod:

• Scheduled: Pod assigned to node
• Pulling: Pulling container image
• Pulled: Successfully pulled image
• Created: Container created
• Started: Container started
• Killing: Container being terminated
• Failed: Container failed to start
• BackOff: Backing off after failed start

ConfigMaps & Secrets

View ConfigMaps and Secrets mounted by this pod:

1. Volume Mounts

   List of volumes mounted into containers:

   • ConfigMap volumes
   • Secret volumes
   • PersistentVolumeClaims
   • EmptyDir volumes
   • HostPath volumes
2. Environment Variables

   Environment variables sourced from ConfigMaps/Secrets:

   • ConfigMapKeyRef
   • SecretKeyRef
   • ConfigMapRef (all keys)
   • SecretRef (all keys)
3. Navigate

   Click ConfigMap/Secret names to view their details

Actions

Restart Pod

Delete the pod to trigger recreation by its controller:

1. Click Restart Button

   In the detail view, click “Restart Pod”

2. Confirm Action

   A confirmation dialog appears

   This deletes the pod. It will be recreated by its Deployment/StatefulSet/DaemonSet controller.

3. Wait for Recreation

   Watch as:

   1. Old pod enters Terminating state
   2. New pod is created (Pending)
   3. Container images are pulled
   4. New pod starts (Running)
4. Verify Health

   Check new pod logs and health probes

When to Restart:

• Container stuck in error state
• Memory leak suspected
• Configuration changes need to be applied
• Debugging intermittent issues
• Force image re-pull

Standalone pods (not managed by a controller) will be permanently deleted

View Logs

Open the logs viewer for detailed log analysis:

1. Select Container

   If multi-container pod, choose which container

2. Set Line Limit

   Choose how many lines to fetch (100-5000)

3. Enable Auto-refresh

   Toggle auto-refresh for live log streaming

4. Search Logs

   Use search bar to filter log lines

5. Download

   Export logs to text file for offline analysis

Troubleshooting

Pod Stuck in Pending

Symptom: Pod remains in Pending state for extended time

Possible Causes:

Insufficient Resources

Cluster doesn’t have enough CPU/memory

Solution:

1. Check events for “Insufficient cpu” or “Insufficient memory”
2. Scale up cluster or reduce resource requests
3. View node capacity in Nodes page

Image Pull Failure

Cannot pull container image

Solution:

1. Check events for image pull errors
2. Verify image name and tag
3. Check registry authentication
4. Test image pull manually: docker pull <image>

Volume Mount Issues

PersistentVolumeClaim not bound or accessible

Solution:

1. Check PVC status
2. Verify StorageClass availability
3. Check volume node affinity

Node Selector Mismatch

No nodes match pod’s nodeSelector/affinity rules

Solution:

1. Check pod’s nodeSelector in YAML
2. Verify node labels match requirements
3. Review affinity/anti-affinity rules

CrashLoopBackOff

Symptom: Container repeatedly crashes and restarts

Diagnostic Steps:

1. Check Container Logs

   View logs for error messages and stack traces

2. Review Exit Code

   Check restart history for exit codes:

   • 137: OOMKilled (increase memory limit)
   • 1/127: Application error (check logs)
   • 139: Segmentation fault (application bug)
3. Verify Configuration

   • Environment variables correct?
   • ConfigMaps/Secrets accessible?
   • Required files/volumes mounted?
4. Check Resource Limits

   • Memory limit sufficient?
   • CPU limit not too restrictive?
   • Increase limits and test
5. Test Locally

   Run container image locally with same configuration:

   docker run -it <image> /bin/sh
   

OOMKilled (Out of Memory)

Symptom: Container killed with exit code 137

Solutions:

1. Increase Memory Limit

   resources:
     limits:
       memory: "1Gi"  # Increase from 512Mi
     requests:
       memory: "512Mi"
   

2. Analyze Memory Usage

   • Check metrics in detail view
   • Profile application memory consumption
   • Look for memory leaks

3. Optimize Application

   • Reduce memory footprint
   • Implement caching strategies
   • Fix memory leaks

ImagePullBackOff

Symptom: Cannot pull container image

Common Solutions:

Authentication

Private Registry Authentication

Create image pull secret:

kubectl create secret docker-registry regcred \
  --docker-server=<registry> \
  --docker-username=<username> \
  --docker-password=<password>

Reference in pod:

imagePullSecrets:
- name: regcred

Image Name

Verify Image Path

Check image name format:

• nginx:latest (Docker Hub)
• gcr.io/project/image:tag (GCR)
• registry.example.com/image:tag (Private)

Common mistakes:

• Wrong registry URL
• Typo in image name
• Non-existent tag

Network

Network Connectivity

Test from node:

# SSH to node
docker pull <image>

Check:

• Internet connectivity
• DNS resolution
• Firewall rules
• Proxy configuration

Failed Readiness Probe

Symptom: Pod running but not ready (not receiving traffic)

Solutions:

1. Check Probe Configuration

   readinessProbe:
     httpGet:
       path: /ready
       port: 8080
     initialDelaySeconds: 10  # Too short?
     periodSeconds: 5
     failureThreshold: 3
   

2. Increase Delays

   • Application needs more time to start
   • Increase initialDelaySeconds
   • Adjust periodSeconds

3. Fix Health Endpoint

   • Ensure /ready endpoint exists
   • Returns 200 status code
   • Responds quickly (< timeout)

4. Check Dependencies

   • Database connection ready?
   • External services accessible?
   • Required files loaded?

Best Practices

Always Set Resource Limits

Prevent resource starvation and OOMKills

resources:
  limits:
    memory: "512Mi"
    cpu: "500m"
  requests:
    memory: "256Mi"
    cpu: "250m"

Configure Health Probes

Enable auto-recovery and traffic management

livenessProbe:
  httpGet:
    path: /healthz
    port: 8080
  initialDelaySeconds: 30
  periodSeconds: 10

readinessProbe:
  httpGet:
    path: /ready
    port: 8080
  initialDelaySeconds: 5
  periodSeconds: 5

Use Specific Image Tags

Avoid latest tag for reproducibility

# Good
image: nginx:1.21.6

# Bad
image: nginx:latest

Implement Graceful Shutdown

Handle SIGTERM for clean termination

lifecycle:
  preStop:
    exec:
      command: ["/bin/sh", "-c", "sleep 15"]

Log to stdout/stderr

Make logs accessible via kubectl/Orphelix

// Good
console.log("Message")

// Bad (logs to file only)
fs.writeFile("/var/log/app.log", "Message")

Related Resources

Deployments

View parent deployment details

Nodes

Check node where pod is running

ConfigMaps & Secrets

View mounted configuration

Events

See all cluster events