Stop 12 · Monitoring & ObservabilityTopics · 7/8

Cost Monitoring & Anomaly Detection

Part of the Cloud Computing Roadmap

In short

A test script stuck in a loop, or a GPU instance you forgot over a long weekend, can cost more than a semester's fees. Cost monitoring catches it in a day, not at month end.

Tracking spend continuously and getting alerted when it deviates unexpectedly from normal patterns — catching a runaway resource or misconfiguration before it turns into a shocking monthly bill.

The blob character looks shocked at a long bill while a geyser glows red behind.

Think of it like this

The geyser left on

The electricity bill arrives at double the usual amount. After some detective work, you find the geyser has been on all month. If the meter had beeped on day two, you'd have saved a lot. Anomaly detection is that beeping meter.

Watch it happen

Daily spend, one odd day

210
0
205
1
220
2
215
3
890
4
900
5
212
6

1/5A week of daily spend in rupees for a small practice account. The normal range is about 200 to 220.

The big ideas

What you need to know

Budgets

Fixed thresholds: alert at 50%, 80% and 100% of a monthly budget, including forecasted spend.

Anomaly detection

AWS Cost Anomaly Detection and similar tools in Azure and Google Cloud flag unusual spend and its likely cause.

Tags and accounts

Separate accounts or projects per environment, plus tags, show who caused a spike.

Billing data lags

Cost data can be hours behind, so pair it with usage alarms for very fast-growing risks.

See it in code

Walk through the code

bash
1aws ce create-anomaly-monitor --anomaly-monitor '{2  "MonitorName": "per-service",3  "MonitorType": "DIMENSIONAL",4  "MonitorDimension": "SERVICE"5}'6 7aws ce get-anomalies --date-interval StartDate=2026-10-01,EndDate=2026-10-27

1/2Create a dimensional monitor that tracks spend per service, like EC2, S3 and Lambda.

A monitor that watches each AWS service's spend separately. Add a subscription in the console to get emailed when it finds something.

Save yourself the pain

Common mistakes

Oops

Checking the bill only when the monthly invoice email arrives.

Do this instead

Set budgets and anomaly alerts on day one, so surprises show up within a day.

Oops

Running dev, test and prod in one account with no tags.

Do this instead

Split accounts or tag everything, so a spike points straight to its owner.

Your mission

Guard your practice account

~20 min

Quick check

Your budget alert is at 100% of ₹2,000. Why also add an alert on forecasted spend?

Go deeper

  1. 1Set up a budget alert in your cloud provider's billing console.
  2. 2Explore a cost anomaly detection tool (many providers offer one built-in).
  3. 3Learn to tag resources by team or project so cost can be attributed and investigated accurately.

More topics at this stop

Stuck on this topic? Ask an Insider

Get 1:1 guidance from people who've walked this exact path — on the InsideEdge app.

Download
InsideEdge