Private early access — opening soon

Kubernetes monitoring
in five minutes.
Not five sprints.

One agent. Zero YAML. No Prometheus, no Alertmanager — just alerts that tell you what broke, before your customers do.

How it works
monidori.com/kubernetes

Cluster name

prod-eu-1
Connected · agent v0.3.3 · prod-eu-1 is reporting

$ kubectl apply -f https://monidori.com/agent/v1/install.yaml

namespace/monidori-system created

serviceaccount/monidori-agent created

deployment.apps/monidori-agent created

$ kubectl -n monidori-system create secret generic monidori-agent --from-literal=token=mdr_••••••

secret/monidori-agent created

01Install the agent

Name the cluster and run the two kubectl commands it gives you. A small read-only agent starts reporting; Monidori never gets your credentials.

Cluster view

Your clusters, without the kubectl.

Nodes, pods, volumes and events in one view, with logs and YAML a click away. No kubectl needed.

NodesPodsWorkloadsStorage

cp-1

Ready

CPU 34%

Memory 52%

worker-1

Ready

CPU 71%

Memory 64%

worker-2

Ready

CPU 46%

Memory 81%

api-6c8b9d-x2kq

api

Running0

checkout-7d9f-mp4s

shop

CrashLoopBackOff14

worker-5f7c-9wqz

jobs

Running1

migrate-28817-zt

shop

Succeeded0

postgres-data shop

91% of 20Gi · full in ~4 days

uploads web

38% of 50Gi

prometheus-db obs

62% of 100Gi

No YAML

Alerting that you actually enjoy.

Skip the PrometheusRules, the Alertmanager routes and the PromQL. Pick a condition, set a threshold, choose a channel — or describe what you want and let AI set it up.

The usual way · two config files

# prometheusrule.yaml
apiVersion: monitoring.coreos.com/v1
kind: PrometheusRule
spec:
  groups:
  - name: pods
    rules:
    - alert: PodCrashLooping
      expr: increase(kube_pod_container_status_restarts_total[15m]) > 5
      for: 5m
# alertmanager.yaml
route:
  receiver: telegram
  group_by: [alertname, namespace]
receivers:
- name: telegram
  telegram_configs: …

With Monidori · one sentence, or four fields

Tell me on Telegram if any pod keeps restarting

Create

Or set it up yourself

WhenPod crash loop
Restarts at least5
NamespaceAll namespaces
NotifyTelegram · Ops
Save rule

Features

One tool instead of four.

Service discovery, cluster visibility, alerting and AI diagnosis — designed together, so they actually talk to each other.

Your services, found and watched

Install the agent and everything running shows up on its own, by namespace. Public sites are checked from outside, so an expiring certificate or a slow page doesn't surprise you.

prod-eu-1

Reading the cluster…

0 services found

api

api · Deployment · 3 replicas

Found

checkout

shop · Deployment · 2 replicas

Found

shop.example.com

shop · Ingress · https

Found

postgres

shop · StatefulSet · 1 replica

Found

billing

billing · Deployment · 2 replicas

New

Deploy something new and it shows up on its own.

billing was just deployed, and it's already here.

AI that reads the cluster for you

Open a cluster and Monidori summarises what's actually wrong: crash-looping pods, unhealthy deployments, warning events and failing CronJobs. Describe the alert you want in plain English and it sets the rule up for you.

AI summary

1 pod is crash-looping: checkout-7d9f in shop restarted 14× after BackOff warnings, and its deployment is 0/2 ready. The nightly-backupCronJob hasn't succeeded since yesterday.

Automatic service discovery
SSL certificate expiry
Pod logs with container picker
Human-readable YAML
No PromQL, no rule files
Disk usage & predictive forecasts
Previous container logs
Namespace include / exclude per rule
Alert presets: Minimal, Default, Debug
OOMKilled, pending, image-pull and node-pressure alerts
CronJob success tracking
K8s version & EOL tracking
Agent-offline alerts
Read-only cluster access
Full alert history

FAQ

Questions, answered.

Who is Monidori for?

Engineers running services on Kubernetes who want to see what’s running and what’s broken without setting up and maintaining a Prometheus + Grafana + Alertmanager stack.

What access does Monidori need to my cluster?

Read-only: the agent can list cluster objects and read pod logs. No Secrets, no changes, outbound HTTPS only. How the agent works →

Do I need Prometheus or Alertmanager?

No. The agent reads the Kubernetes API directly: no exporters, no rule files, no Alertmanager routes.

Does it replace Prometheus or Datadog?

Not for custom metrics, PromQL or tracing. Monidori answers “is it up, and what broke?”. Run Prometheus alongside it if you need more.

How fresh is the cluster data?

Updated every 30 seconds, or right away with Refresh now. Logs and YAML are read live when you open them.

When can I sign up?

Not yet: Monidori is in private early access. Check back soon.

Stop finding out from your users.

Monidori is in private early access. Signups open soon.