704.4 Tracing (weight: 2) DEVOPS-TOOLS-ENGINEER Practice Questions (Page 2)

Questions 6–10

1. 6application · medium

   A company is standardizing its observability stack. They use Prometheus for metrics and Grafana for dashboards. They are introducing distributed tracing and want a tool that integrates well with their existing Kubernetes environment, supports OpenTelemetry natively, and offers a UI for trace analysis. They do not want to manage a separate database for trace storage. Which tool should they choose?

   • AZipkin with the in-memory storage backend.
   • BGrafana Tempo, which uses object storage for traces and integrates with Grafana.
   • COpenTelemetry Collector alone, without a backend.
   • DJaeger with the Elasticsearch storage backend.
   Reveal Solution Discussion Select an answer first
2. 7application · medium

   A developer is instrumenting a microservice that receives a request from an upstream service. The upstream service uses OpenTelemetry and injects a `traceparent` header. The developer needs to ensure the new service continues the same trace. What should the developer do in the service's request handler?

   • AExtract the `traceparent` header and use it to create a child span in the current trace.
   • BCreate a new trace context and generate a new trace ID for the incoming request.
   • CLog the header value and manually search for it in the tracing backend.
   • DIgnore the header and rely on the service's own logging to correlate the request.
   Reveal Solution Discussion Select an answer first
4. 8application · medium

   A company has a microservices architecture. They are experiencing intermittent latency issues that are difficult to diagnose because the problem spans multiple services. They already have logging and metrics in place. Which additional practice would most directly help them understand the flow of a single request across all services?

   • AIncrease the frequency of metric collection to every second.
   • BAdd more detailed logging to every service to capture the exact time of each operation.
   • CImplement distributed tracing to capture the end-to-end path of a request as it moves through services.
   • DUse a network monitoring tool to analyze traffic between services.
   Reveal Solution Discussion Select an answer first
5. 9application · medium

   A developer is adding tracing to a service that communicates with a legacy system over a proprietary binary protocol. The protocol does not support HTTP headers. How can the developer propagate trace context to the legacy system?

   • AStart a new trace for the legacy system call.
   • BInclude the trace context in the payload of the binary protocol.
   • CLog the trace ID and manually correlate it in the tracing backend.
   • DUse a separate HTTP call to the legacy system just to send the trace context.
   Reveal Solution Discussion Select an answer first
6. 10expert · hard

   A team is using Jaeger with head-based sampling at 5% to control costs. They are now debugging a rare issue that only occurs when a specific combination of services is called in sequence. They suspect the issue is in a service that is only called in this rare path. They want to ensure they capture this rare path without significantly increasing storage costs. What is the best approach?

   • ADisable sampling for all services involved in the rare path.
   • BIncrease the sampling rate to 100% for the first service in the sequence.
   • CUse a fixed sampling rate of 50% for all services.
   • DUse tail-based sampling with a policy to keep all traces that involve the specific service in the rare path.
   Reveal Solution Discussion Select an answer first