3.1 - Cloud monitoring and observability
Cloud+ CV0-004 objective 3.1 covers monitoring and observability in the cloud. Metrics are numeric time-series values such as CPU percent and request rate, collected for dashboards and analysis. Alerting sends a notification when a metric crosses a defined threshold, for example when CPU usage exceeds a limit, so operators react quickly. Log aggregation collects logs from many services into one searchable platform, making cross-service troubleshooting practical. An SLO (service level objective) is a measurable target, such as 99.9% uptime, that the service commits to meeting, and it underpins error budgets and SLAs. You should also know traces, the three pillars of observability (metrics, logs, traces), and dashboards. Expect scenario questions that describe collecting time-series data, notifying on a threshold, centralising logs or setting a reliability target, and ask which concept - metrics, alerting, log aggregation or SLO - applies.
Numeric time-series (CPU %, request rate) = metrics. Notify when a threshold is crossed = alerting. Many services' logs in one searchable place = log aggregation. Measurable reliability target (99.9%) = SLO.
Practice questions
1. Sending a notification when CPU usage crosses a defined threshold is an example of:
- Logging
- Alerting (correct answer)
- Tracing
- Indexing
Alerting fires notifications when a metric breaches a threshold. Logging records events; tracing follows a request across services.
2. Collecting logs from many services into one searchable platform is called:
- Log rotation
- Log aggregation (correct answer)
- Log masking
- Log signing
Log aggregation centralizes logs from many sources into a searchable store (ELK, Loki), enabling correlation and troubleshooting. Rotation just recycles old files.
3. Following a request as it passes through several microservices to find where it slows down uses:
- Metrics aggregation
- Centralized logging
- Distributed tracing (correct answer)
- Threshold alerting
Distributed tracing follows one request across services with span timings, exposing the slow hop. Metrics show aggregate numbers; logs record discrete events.
4. A measurable target such as 99.9% uptime that the service commits to meeting is a:
- SLI
- SLO (correct answer)
- KPI
- MTTR
An SLO (service level objective) is the internal target for a metric, like 99.9% uptime. An SLI is the measured value; an SLA is the external contract with penalties.
5. Setting an alert only when high CPU lasts several minutes, not on a brief spike, mainly reduces:
- Alert noise (correct answer)
- Storage cost
- Network latency
- Backup size
Requiring a threshold breach to persist for a duration filters out transient spikes, cutting false alarms and alert fatigue. It does not affect storage, latency or backups.
6. Numeric time-series values like CPU percent and request rate collected for dashboards are:
- Traces
- Metrics (correct answer)
- Logs
- Events
Metrics are numeric time-series measurements aggregated for trending and alerting. Logs are discrete text records of events; traces follow a single request across services.