Management, Monitoring, and Billing
Domain 6 - Management, Monitoring, and Billing. This domain tests observability tooling and how cost is actually metered, which is where surprise invoices come from.
Monitoring and operations
- Cloud Insight: metrics, dashboards, alarms and notifications; the server monitoring agent must be installed and running, and alarms need a notification channel.
- Cloud Log Analytics: collects and searches system and application logs; retention is a setting, not infinite.
- Cloud Activity Tracer: audit of console and API actions per account.
- Resource Manager: template-based provisioning for repeatable environments.
Billing mechanics
Charging is usage-based: Server by spec and running time, Block Storage by allocated GB regardless of use, Public IP while allocated even when detached, outbound traffic, managed database by spec plus storage, AI APIs per call or character. Prepaid credit and budget alerts control spend; separate accounts per environment keep cost attributable.
Worked example
The monthly bill triples. Reasoning: use Cloud Activity Tracer to find recently created resources and who created them; check Cloud Insight for idle but running servers; look for detached Public IPs, oversized Block Storage and stale snapshots; confirm OCR or AI call volume; then set a budget alert and tag resources by owner.
Traps
- Assuming a stopped server is free - storage and reserved IPs still bill.
- Ignoring egress traffic, often the surprise line item.
- Expecting metrics without the monitoring agent installed.
- Assuming logs are retained forever.
- Confusing a monitoring alarm with a scaling policy: an alarm must be linked to a scaling action to resize anything.