DevOps Monitoring & Observability 2 — Questions and Answers
Question 1: What is distributed tracing?
- Tracking a request's journey across multiple services to diagnose latency and failures (Correct answer)
- Monitoring CPU and memory usage across distributed servers
- Logging all API calls made by a distributed application
- Collecting metrics from multiple cloud providers in one place
Correct answer: Tracking a request's journey across multiple services to diagnose latency and failures
Distributed tracing instruments requests with trace IDs that flow through all microservices, allowing developers to visualize the full request path and identify bottlenecks.
Question 2: What is the ELK Stack?
- Elasticsearch, Logstash, and Kibana — a log aggregation, search, and visualization stack (Correct answer)
- Event Logger, Load balancer, and Kubernetes — a cloud-native observability stack
- Elastic, Linkerd, and Kong — a service mesh and API gateway combination
- Event-driven Logging with Kafka — a streaming log pipeline
Correct answer: Elasticsearch, Logstash, and Kibana — a log aggregation, search, and visualization stack
The ELK Stack combines Elasticsearch (search/storage), Logstash (log ingestion/transformation), and Kibana (visualization) to build a centralized logging platform.
Question 3: What is a 'golden signal' in SRE monitoring?
- One of four key metrics: latency, traffic, errors, and saturation (Correct answer)
- An alert threshold that triggers a P1 incident response
- A baseline metric measured during system initialization
- A synthetic monitoring check run from external locations
Correct answer: One of four key metrics: latency, traffic, errors, and saturation
Google SRE defined four golden signals — latency, traffic, errors, and saturation — as the most critical metrics to monitor for any service.
Question 4: What is alerting 'noise' in a monitoring context?
- Excessive false-positive alerts that distract on-call engineers from real issues (Correct answer)
- Background network traffic that interferes with monitoring data collection
- Duplicate metrics collected from redundant monitoring agents
- Log verbosity that obscures important error messages
Correct answer: Excessive false-positive alerts that distract on-call engineers from real issues
Alert noise refers to too many low-priority or false-positive alerts that cause alert fatigue, leading engineers to ignore or silence important notifications.
Question 5: What is OpenTelemetry?
- A vendor-neutral observability framework providing APIs, SDKs, and tools for logs, metrics, and traces (Correct answer)
- An open-source alternative to Prometheus for metrics collection
- A cloud-native logging standard for containerized applications
- A distributed tracing protocol developed by Google
Correct answer: A vendor-neutral observability framework providing APIs, SDKs, and tools for logs, metrics, and traces
OpenTelemetry is a CNCF project that standardizes telemetry data collection so applications can be instrumented once and send data to any compatible backend.
Question 6: What is a 'runbook' in incident management?
- A documented set of procedures for responding to specific operational events or alerts (Correct answer)
- A record of all incidents that occurred during a given period
- An automated script that executes remediation steps without human intervention
- A dashboard showing the real-time health of production systems
Correct answer: A documented set of procedures for responding to specific operational events or alerts
A runbook provides step-by-step instructions for handling a specific alert or operational task, reducing response time and cognitive load during incidents.
DevOps Monitoring & Observability 2
What is distributed tracing?