app.kubewatchlabs.com / overview
Live
Cluster throughput live
48.2k-10.1%req/min
Requests Errors
Resource usage6 nodes
CPU34%
Memory61%
Disk78%
Latency by servicep95 · ms
authquerylivealertbilltenant
Active alerts2 firing
API latency p99 > 800msgateway2m
Memory pressure on node-3kubelet14m
Pod restart loop clearedbilling1h
Disk usage back under 80%postgres3h
Containers128 total
gateway-7f4c
running
auth-9b2a
running
query-3d8e
running
billing-1a5f
stopped
Cloud spendthis month
$3,2840.5%
Zero-config agent · Kubernetes + Docker

Monitor every container
from one dashboard

Deploy a lightweight agent in 60 seconds. No cloud credentials required. Get full visibility into your infrastructure, CPU, memory, logs, and alerts, in minutes.

Start your trial, no card required
60s
To first metric
< 0.5%
Agent CPU overhead
99.9%
Uptime SLA

Integrates with the tools you already run

PostgreSQL
PostgreSQL
MySQL
MySQL
Redis
Redis
Kafka
Kafka
Argo CD
Argo CD
Jenkins
Jenkins
Prometheus
Prometheus
Grafana
Grafana
Netlify
Netlify
Vercel
Vercel
Cloudflare
Cloudflare
Snowflake
Snowflake
Databricks
Databricks
MongoDB Atlas
MongoDB Atlas
Jira
Jira
Gitea
Gitea
Vault
Vault
Elasticsearch
Elasticsearch
RabbitMQ
RabbitMQ
NGINX
NGINX
Istio
Istio
CockroachDB
CockroachDB
InfluxDB
InfluxDB
MinIO
MinIO
Capabilities

Everything you need to stay on top of your containers

From single Docker hosts to multi-cluster Kubernetes environments.

Docker Monitoring

Full visibility into every container, CPU, memory, network I/O, live log streaming, and restart tracking.

  • Container list & status
  • CPU & memory metrics
  • Network I/O per container
  • Live log tail
  • Image digests & volume mounts
  • Interactive exec & volume browsingEnterprise

Kubernetes Monitoring

Understand your cluster at a glance. Connect a cluster in one click or bring your own kubeconfig, then track pods, nodes, services, and per-node resource usage without leaving your browser.

  • Auto-discover EKS, GKE & AKS clusters
  • Per-node CPU, memory, network & disk IOPS
  • API server request rate
  • Classic Ingress and Gateway API side by side
  • Read-write cluster tools with self-service permission togglesEnterprise

Cluster & Host Backups

Back up the actual infrastructure your agents watch, not just KubeWatch itself: Kubernetes object manifests by namespace or label selector, or named volumes and bind mounts on a Docker host.

  • Scheduled or on-demand, per agent
  • S3, SFTP, webhook & Azure Blob destinations
  • Per-namespace / per-mount result history
  • Bytes stream straight from the agent, never through our servers

Node Cordon & Drain

Move workloads off a node before maintenance, a resize, or decommissioning it, respecting every PodDisruptionBudget you have configured.

  • Cordon, drain & uncordon from the Nodes page
  • PDB-aware eviction with automatic backoff
  • DaemonSet pods & emptyDir data called out before you approve
  • Admin-approved before anything is applied

Vulnerability Scanning

Scan every container image for known CVEs with your choice of engine, Trivy, Grype, or Snyk, surfaced right on the image detail page next to its digest and size.

  • Trivy, Grype & Snyk support
  • Severity-ranked CVE findings
  • Scan on demand, cached per image digest
  • Reverse lookup: every affected container & pod

Real-time Alerts

Know before your users do. Set threshold rules, route to Slack or email, and silence noisy alerts intelligently.

  • Threshold-based rules
  • Slack & email channels
  • Silence detection
  • Alert history & audit log

Incident Management

Detect, respond, resolve, and learn, all in one place. Escalate a firing alert into an incident, work it with your team, and capture a postmortem when it is done.

  • On-call scheduling & dynamic routing
  • Timeline, responders & related incidents
  • Postmortems with action items
  • MTTA / MTTR insights

Auto Scaling

Close the loop from watching to acting. Scale Kubernetes pods and Docker containers on the metrics you already track, from one policy.

  • Native HPA & Karpenter on K8s
  • Orchestrated scaling on Docker
  • Dry-run, cooldowns & approvals
  • One-click rollback

Integrations

Monitor the services around your apps. Connect databases, caches, message queues, CI/CD, and observability tools with live health and deep metrics.

  • Postgres, MySQL, Redis, Kafka
  • Argo CD, Jenkins, Prometheus, Grafana
  • Latency & uptime health checks
  • Per-service deep metrics

Performance & Load Testing

Run performance, load, stress, soak, and spike tests against any endpoint and read real latency percentiles, throughput, and a full success and error breakdown.

  • 5 test types: load, stress, soak, spike & more
  • P50 / P90 / P99 / max latency
  • Breaking point & recovery time
  • Uptime & latency-drift tracking

Kafka Topics Monitoring

Watch topic size and consumer group health without a separate collector. The agent talks directly to your Kafka cluster's admin API, filterable by cluster, topic, and consumer group.

  • Topic partition offsets over time
  • Consumer group total & per-partition lag
  • Committed offset tracking
  • No consuming, no separate service

AI Agent Cost & Reliability Ops

Bring your AI agents (LangChain, CrewAI, custom OTel-instrumented systems) into KubeWatch as a first-class monitored entity, with the same alerting your infrastructure already gets.

  • Per-session cost, tokens & tool calls
  • Waterfall trace view per session
  • Cost, error-rate & runaway-session policy rules
  • Dry-run policy evaluation with a full decision log

Docker-vs-Kubernetes Advisor

A workload-level recommendation, backed by its own historical usage, of whether it needs Kubernetes or would run just as well as a KubeWatch-managed Docker replica.

  • CPU volatility & restart-pattern scoring
  • Structured, readable rationale per recommendation
  • Conservative savings estimates
  • Advisory only, nothing applied automatically

GPU & Inference Cost Control

The same autoscaler policy engine, made GPU- and cost-aware. Live GPU utilization, cost-per-1k-tokens attribution, and GPU-aware scaling policies.

  • Real-time GPU utilization, memory & power
  • Cost-per-1k-tokens attribution
  • Scale-to-zero, rightsize & spot-shift policies
  • Same dry-run & rollback conventions

Self-Hosted AI Stack Observability

Treat vLLM, Ollama, SGLang, and TGI as first-class monitored workloads, with a Model Registry and a Sovereign Fallback Health Checker that proves your fallback path actually works.

  • vLLM / Ollama / SGLang / TGI metrics
  • Model Registry by role & version
  • Sovereign Fallback Health Checker
  • Agent-side probing, read-only

Autonomous Remediation

Attach an automated restart to any alert rule you already have. Every playbook starts in forced dry-run, is rate-limited by a circuit breaker, and defaults to requiring human approval.

  • Docker container restart & K8s rolling restart
  • Forced dry-run until you promote it
  • Per-playbook circuit breaker
  • Full append-only audit trail

Edge AI Fleet Observability

Bring cameras, sensors, and industrial controllers into the same fleet view as your containers and nodes, over the same long-poll channel built for the autoscaler.

  • Store-and-forward buffering for spotty links
  • Three-state connectivity: online / expected / unexpected
  • Per-device offline windows, not one global grace period
  • OTA model updates via your own apply script

Infrastructure Automation

A GitOps control plane for OpenTofu and Ansible: sync Projects from Git, run Stacks through a plan-then-approve-then-apply workflow, and inject secrets from an encrypted vault, never plaintext, never committed to your repo.

  • Git-synced Projects & Stacks, OpenTofu or Ansible
  • Mandatory human-approved dry run before any real change
  • AES-256 encrypted secrets vault, including SSH keys
  • 50 built-in blueprints to scaffold a new Stack straight to Git
  • Project-scoped roles & a full audit trail

DataOps

Orchestrate and monitor your data pipelines and cloud infrastructure, all from one place. Trigger runs across AWS Glue/Step Functions, Airflow, Azure Data Factory, and GCP Dataflow/Composer, then start, stop, resize, or delete databases, storage, and serverless functions across AWS, Azure, and GCP, without leaving KubeWatch.

  • Trigger, cancel & monitor runs across 4 ETL tool families
  • Automatic run history, refreshed roughly every 20 seconds
  • Database, storage & serverless inventory across AWS, Azure & GCP
  • Start, stop, resize & delete resources with a confirm step
  • Admin-only, available on every plan

Analytics

Connect your own Postgres, MySQL, Redshift, or S3 data, clean it with a pipeline that runs natively inside KubeWatch, and chart the result right on the dashboard, no external BI tool needed.

  • Postgres, MySQL, Redshift & S3 connections
  • Encrypted-at-rest credentials
  • Dedupe, fill/drop missing, coerce, filter & custom-transform steps
  • Manual or recurring-interval runs
  • Line, bar & area charts from a cleaned dataset
  • Admin-only, available on every plan
Setup

Up and running in minutes

No complex setup. No cloud IAM roles. Just deploy and watch.

01

Sign up in 30 seconds

Create your account with just your email. No credit card, no sales call, no waiting.

02

Deploy the agent with one command

Run a single Docker or Helm command on your host. The agent securely streams metrics to your dashboard.

03

See your infrastructure in under 5 minutes

Within minutes you have a live view of every container and node, CPU, memory, logs, and more.

Deployment

Two ways to run KubeWatch

Choose the deployment model that fits your team.

Hosted SaaS

30-day trial, no card required. We manage the infrastructure.

  • Agent-only deployment
  • We manage the backend
  • No credit card to start
  • Scale with your team
Start trial

Self-Hosted

Pro or Enterprise. Your data stays in your infrastructure.

  • Full platform deployment
  • Data residency & compliance
  • Docker Compose or Helm install
  • Annual license with updates
Learn more
Pricing

Simple, transparent pricing

Start with a 30-day trial. Choose Pro or Enterprise before it ends.

Most popular

Pro

$20/mo or $400/yr
  • 50 agents
  • 10 users
  • 50 alert rules
Start trial

Enterprise

$40/mo or $600/yr
  • 200 agents
  • Unlimited users
  • Unlimited alert rules
Start trial
View full pricing →
30-day Enterprise trial · No credit card

See your whole stack in under 5 minutes

Join engineering teams who replaced their scattered dashboards with one place for Docker and Kubernetes.

Start your trialCompare plans