🧐 What is kwatch?

kwatch is like a smart friend for your Kubernetes cluster:

💥

Something crashes → you get a message that says why (not just "pod is broken")

🔇

Smart about noise — groups related issues, ignores flapping, sends a digest when things get crazy

🧠

Explains itself — every alert names the cause, the impact, and what recently changed

⚡

Works in under a minute — just one command and a config file

🚫 No Prometheus🚫 No Grafana🚫 No 50-step setup
Just alerts that make sense. 🎯

🆚 kwatch vs the scary stuff

Feature	👑 kwatch	😰 DIY	💸 SaaS
⏱️ Setup time	~5 minutes	hours of YAML	agent + backend setup
📦 Size	~20 MB single binary	whole monitoring stack	per-node agents + cloud costs
💬 Alerts	Self-explaining ("OOMKilled — raise memory limit")	Rule-defined message	Depends on configuration
🗄️ Storage	None (stateless)	Prometheus TSDB	Full retention (costly)
📚 Learning curve	One ConfigMap	PromQL + alert rules	Platform-specific DSL

🚨 Before vs After

🤷 Before

CrashLoopBackOff

😰 Raw YAML output — good luck figuring it out

Error

😰 Just "Error" — no context, no cause

⬌kwatch

💡 After

🚨 OOMKilled

💡 memory limit: 512Mi — try raising limits.memory

📋 logs + events included

🚨 HTTP probe failing

💡 on :8080/healthz (exit 137) — container ran out of memory

🔧 suggests raising memory limit

🎯 What does it catch?

Every monitor below is on by default — zero config needed:

✅🟥 Pod crashes (CrashLoop, OOM, ImagePull, Error)

Container state + last logs + events — tells you *why*

✅⏳ Pending pods (stuck Unschedulable)

Alerts after 300s stuck

✅🖥️ Node issues (NotReady, Disk/Memory pressure)

Per-condition severity

✅💾 PVC running out of space

Warn at 80%, critical at 90%

✅❌ Failed Jobs & stuck CronJobs

JobFailed / suspended / missed runs

✅🚀 Stuck rollouts & StatefulSets

ProgressDeadlineExceeded — deployment didn't finish

✅📡 DaemonSet pods not running

Unavailable pods detected

✅📈 HPA stuck at max replicas

After 20 minutes sustained

✅📣 Cluster autoscaler can't scale

FailedToScaleUp / NotTriggerScaleUp

✅🔒 TLS certs expiring

Enable if you want cert expiry warnings

✅💓 Heartbeat (dead man's switch)

Enable to page you if kwatch itself goes down

✅ TLS and heartbeat are the only ones off — everything else just works out of the box.

🧠 Alerts that explain themselves

kwatch ships with a diagnosis engine that runs inside your cluster, reads the logs and events, and tells you what's wrong and what to do next.

How does the insight engine work?

Something breaks — a pod crashes, a node goes down, a deployment gets stuck

kwatch works out the root cause — it maps the pod to its node, owner, services, PVCs and config, and reads the container logs, events, and crash reasons

You get a plain-English fix — "OOMKilled — try raising memory limit" instead of a cryptic error code

🕳️ Knows when it was blind: kwatch stamps its own liveness, so if it was down while your cluster wasn't, the next startup message says how long nobody was watching.

diagnosis configuration

# Says *why* a crash happened — the root cause, impact, and what changed.
# Configuration is optional: the insight/dependency graph is on by default.

When a crash happens, the diagnosis engine reads the logs and tells you the most likely cause and what to do next. Like having a senior SRE on-call with you. 🎯

⚡ Installation

60-second install

Pick your method — both take under a minute 🚀

🏆Easiest

📦 Helm

helm repo add kwatch https://kwatch.dev/charts
helm install [RELEASE_NAME] kwatch/kwatch --namespace kwatch --create-namespace --version 0.11.0-rc.6

More details in the chart docs 📖

🐙Classic

kubectl

Get config

curl -L https://raw.githubusercontent.com/abahmed/kwatch/v0.11.0-rc.6/deploy/config.yaml -o config.yaml

Edit & apply

vim config.yaml  # ✏️ add your webhook
kubectl apply -f config.yaml

Deploy kwatch 🎉

kubectl apply -f https://raw.githubusercontent.com/abahmed/kwatch/v0.11.0-rc.6/deploy/deploy.yaml

📖 Not a monitoring platform — and proud of it! 🎉

kwatch is not a metrics collector, dashboard, or observability backend. No TSDB, no dashboards, no log storage, no query language. kwatch is the alarm — your existing tools are the archive.

Need full observability? Pair kwatch with Prometheus + Grafana for metrics, or Loki for logs. kwatch handles the one thing a dashboard cannot: telling you something broke right now. ⏰