How is an anomaly detected?

Each day of spend is compared with the median of the prior 14 days, and must exceed both an absolute and a percentage increase, so small or noisy lines do not page anyone.

Who receives the alert?

The owner of the team, cluster, account or service it belongs to, over Slack, Microsoft Teams, email or a webhook, so it lands with someone who can act.

Can we ask why a cost moved?

Yes. Ask KubeOn answers from your own cost data and shows the figures behind each answer, and every figure is checked against the data before it is shown.

Budgets and anomalies

Catch a spike at the next billing run

Budgets on allocated cost with a run-rate forecast, and anomaly detection against a 14-day baseline, delivered to Slack, Microsoft Teams, email or your on-call tool.

Book a demo Platform overview

Budgets & forecasts

Budgets on any slice of spend, checked every day against actuals and the month-end forecast.

New budget

Acme total

Monthly · All spend

On track

$294,190

of $1,040,000 · 28.3% used

$1,013,322

forecast · 97%

AI and LLM spend

Monthly · ServiceCategory: AI and Machine Learning

Forecast over

$49,732

of $165,000 · 30.1% used

$171,300

forecast · 104%

Digital Commerce

Monthly · Business unit: Digital Commerce

On track

$123,459

of $440,000 · 28.1% used

$425,246

forecast · 97%

AWS

Monthly · ProviderName: AWS

On track

$198,284

of $700,000 · 28.3% used

$682,977

forecast · 98%

What UnitVane brings to this job

Know before month end, not after the invoice

Budgets and anomalies

Budgets on any slice of spend with a weekday-aware month-end forecast. Anomalies are scored against a 28-day median with the resources that drove them. Alerts fire once per threshold and resolve once.

Explore Budgets and anomalies

Budgets & forecasts

Budgets on any slice of spend, checked every day against actuals and the month-end forecast.

New budget

Acme total

Monthly · All spend

On track

$294,190

of $1,040,000 · 28.3% used

$1,013,322

forecast · 97%

AI and LLM spend

Monthly · ServiceCategory: AI and Machine Learning

Forecast over

$49,732

of $165,000 · 30.1% used

$171,300

forecast · 104%

Digital Commerce

Monthly · Business unit: Digital Commerce

On track

$123,459

of $440,000 · 28.1% used

$425,246

forecast · 97%

AWS

Monthly · ProviderName: AWS

On track

$198,284

of $700,000 · 28.3% used

$682,977

forecast · 98%

Budgets, alerts and anomalies on allocated cost

Governance

Set budgets by cluster, namespace, team or account with a 7-day run-rate forecast. Anomalies are flagged against a 14-day median and sent to Slack, Microsoft Teams, email or a webhook.

Explore Governance

Budgets

Data readyUpdated 12 min ago

Last 7 days▾

Budgets

Spend limits for clusters, accounts, teams or any label and tag, across Kubernetes and the cloud bill, with threshold and forecast notifications.

Budgets

4

Over budget

1

Forecast to overspend

2

Monthly spent / budgeted

$27.7K / $81.5K

4 budgets · October 2026, month to date Filter budgets

Budget	Period	Budget	Spent	Forecast	Status
Payments platform Kubernetes · team: Payments	Monthly2026-10	$48,000	$15,14231.5%	$52,203108.8%	Forecast over
Data platform Kubernetes · namespace: data-pipeline	Monthly2026-10	$14,000	$3,60125.7%	$13,07393.4%	OK
Staging clusters Kubernetes · environment: staging	Monthly2026-10	$4,000	$4,374109.3%	$15,093377.3%	Over budget
ML inference Kubernetes · namespace: ml-inference	Monthly2026-10	$15,500	$4,61129.7%	$16,006103.3%	Forecast over

AI spend with an owner, a budget and an alert

AI spend control

Attribute AI spend to product, team and tenant from tags and gateway keys, set budgets with alerts at an amount or a share of budget, and build drill-down reports across every cloud.

Explore AI spend control

Budgets

Data readyUpdated 12 min ago

Last 7 days▾

AI budgets and alerts

By product, team, tenant or deployment, at an amount or a share of budget

Scope	Budget	Spent	Status
Product: Customer supportSupport assistant	$60,000	30.4%	Forecast over
Team: SearchSearch	$45,000	24.5%	OK
Tenant: EU customersPlatform AI	$25,000	33.8%	Forecast over
Deployment: batch-extractDocument processing	15% of tenant	9%	OK

Alerts go to the owner by email, Microsoft Teams or Slack. Gateway keys can stop at 100%.

Answers about spend with every figure traced

Ask KubeOn

Ask about spend, idle, budgets, anomalies or savings. The assistant calls 10 read-only tools under your permissions, and every amount and percentage in an answer is traced to tool output.

Explore Ask KubeOn

Anomalies

Ask KubeOn

What drove this spike?

Checked 3 data sources

Check cost anomalies
Break down Kubernetes cost
Compare with the previous period

Spend in data-pipeline rose to $1,014.60 on 10-07, against a 14-day median of $293.30. Almost all of it came from one workload:

Workload	Before	After
spark-etl-nightly	$82.04	$781.24
kafka-connect	$72.48	$78.02
dbt-runner	$37.66	$40.32

spark-etl-nightly ran 6× more executors after a config change. See it in Allocations.

All 8 figures traced to KubeOn data

e.g. Why did prod rise this week?

Read-only: it can explain and point to changes, never make them. 12 of 50 questions today.

The problem

Why it is hard today

Forecasts and spikes, sent to the owner

Surprises arrive with the invoice:
By the time the bill shows an increase, the spend has run for weeks.
Alerts nobody owns:
An account-level alert reaches a shared inbox, not the team that caused it.
Noisy thresholds:
Fixed limits fire on every busy day and get muted.

How UnitVane handles it

What you get

Run-rate forecasts

Spend to date plus 7 days of run rate for the rest of the period.

Median-based anomaly scoring

$50 and 30% over the 14-day median, and 3x typical variation.

Alert rules

Spend, growth, efficiency, anomalies and agent health.

Owner routing

Budgets per team, namespace, cluster or account.

Hourly evaluation

Checked every hour, notified once per threshold.

Explainable

Every anomaly shows spend, expected value and score.

Outcomes

What changes for your team

Hours, not weeks

Spikes are scored on the next billing run and sent to the owner.

Alerts people keep on

Three thresholds at once keep false alarms down.

Your first weeks

What you see, and when

From the first agent run to findings that need a full usage history.

01
Within the first hour
The cluster, its nodes, namespaces and workloads appear in Diagnostics. Allocated cost per namespace, team and label follows as soon as billing data is in.
02
Once billing data is in
180 days of history are read and allocated, plus 30 days of per-pod split cost history where AWS split cost allocation data exists. A brand-new AWS export can take up to a day to deliver its first file. The allocated history is kept; the raw billing rows leave the hub after 90 days.
03
Within 30 minutes of the first allocation
Savings are computed, starting with findings that need no usage history, such as released volumes that still bill. Change proposals follow at the next daily regeneration.
04
After 7 days of usage
Container requests are right-sized from p95 usage over 7 days of hourly samples.
05
After 14 days
Node group sizing and abandoned workloads, which each need 14 days of data, and over-provisioned volumes, which also need Prometheus volume usage.

UnitVane works with your technology

Every major cloud, any Kubernetes, and the tools and model providers your teams already use.

All integrations

AWS

Microsoft Azure

Google Cloud

Cloud

Kubernetes

Red Hat OpenShift

Rancher

Tanzu

k3s

Helm

Terraform

Argo CD

Flux

Prometheus

OpenTelemetry

Okta

Slack

Microsoft Teams

OpenAI

Anthropic

Mistral AI

Your questions answered

Keep reading

Expert guides and related solutions

Article

Where Kubernetes spend goes: what the public data says

Independent surveys and billing data agree: most of what you pay for in a Kubernetes cluster is capacity no workload uses. Where it goes, why Kubernetes makes it worse, and the order to fix it in.

Article

Chargeback that survives a reorganization

Chargeback works for a few months, then a reorganization silently rewrites every past month. How to give ownership a date, let labels do most of the work, and close each month on numbers that never move.

Solution by use case

Showback and chargeback

A bill for every team, idle and shared included

Solution by use case

Kubernetes rightsizing

Requests and nodes sized from measured usage

Solution by use case

Idle and waste cleanup

Abandoned workloads and orphaned volumes, found and priced

All resources

Ready to see it on your own data?

A 30-minute walkthrough on your own billing data, for your clouds, your clusters and your model providers.

Book a demo Plan an evaluation

Catch a spike at the next billing run

What UnitVane brings to this job

Budgets and anomalies

Governance

AI spend control

Ask KubeOn

Why it is hard today

Surprises arrive with the invoice:

Alerts nobody owns:

Noisy thresholds:

What you get

Run-rate forecasts

Median-based anomaly scoring

Alert rules

Owner routing

Hourly evaluation

Explainable

What changes for your team

Hours, not weeks

Alerts people keep on

What you see, and when

Within the first hour

Once billing data is in

Within 30 minutes of the first allocation

After 7 days of usage

After 14 days

Your questions answered

How is an anomaly detected?

Who receives the alert?

Can we ask why a cost moved?

Expert guides and related solutions

Ready to see it on your own data?