Skip to content
Tutorials Center
My learning

Topics

Lesson 37 of 37 ยท Projects and Practice

Practice Questions and Next Steps

Test what you learned about Datadog APM, monitors, SLOs, synthetics, security and cost with a practice quiz, a cheat sheet, a query drill and a learning path.

You have covered the whole path from tracing to incident response. This final lesson gives you practice material, a cheat sheet and ideas for what to learn next. As with the rest of the track, Datadog itself cannot run on this site, so the exercises are paper-and-code drills that use runnable stand-ins.

Cheat sheet

Need Datadog feature Remember
Find where a request is slow APM traces, flame graph Self time shows the real bottleneck
Connect traces, logs, metrics DD_SERVICE, DD_ENV, DD_VERSION, DD_LOGS_INJECTION Unified service tagging
Watch a number Metric monitor Time window, scope, thresholds, recovery
Reduce false alarms Composite, anomaly, recovery thresholds, evaluation delay Alert on symptoms
Promise reliability SLO, error budget, burn rate Fast and slow windows
Check from outside Synthetic API and browser tests Several locations, retries
Measure real users RUM, Core Web Vitals p75, source maps, privacy
Find security issues Cloud SIEM, App and API Protection, Cloud Security Tune before you expand
Control cost Metrics without Limits, exclusion filters, archives No IDs in tags
Manage as code API, Terraform, datadog-ci Keys in secrets
Respond Incident Management, On-Call Mitigate first, blameless postmortem

Drill 1: read a monitor query

For each query, say what it measures and when it fires.

TEXT
avg(last_10m):avg:system.mem.pct_usable{env:prod} by {host} < 0.1
sum(last_5m):sum:trace.http.request.errors{service:shop-api}.as_count() > 50
change(avg(last_1h),last_1d):avg:orders.created{env:prod} < -0.5

Answers: (1) memory with less than 10% usable averaged over ten minutes, one alert per host; (2) more than 50 errors in five minutes for the service; (3) the hourly average of orders dropped by more than 50% compared to the same hour one day earlier (a change monitor; the negative threshold means a decrease).

Drill 2: score your own answers

This stand-in grades a short list of statements. Edit the answers object to check yourself (true or false), then run it.

JS Self-check: true or false

Press Run (or Ctrl + Enter) to see the console output here.

No output.

Drill 3: pick the feature

JS Stand-in: match problems to features

Press Run (or Ctrl + Enter) to see the console output here.

No output.

Cover the right column and answer before you run it.

Practice quiz

Quick check ยท 5 questions

of 5 answered You scored / 5 ยท perfect!

  1. 1. A trace shows a 600 ms root span whose only child takes 590 ms. Where should you look first?

    The root span has only 10 ms of self time; the child owns the time.
  2. 2. You want to alert on the error rate of a service. Which source is most reliable under sampling?

    Trace metrics come from all traffic before sampling.
  3. 3. A monitor flaps between Alert and OK every few minutes around 90% CPU. What is the best fix?

    Recovery thresholds and longer windows stabilise the state.
  4. 4. Your custom metric has tags region (5), plan (4) and endpoint (50). What is the worst-case number of series?

    5 x 4 x 50 = 1,000.
  5. 5. Which key should never appear in frontend code?

    API keys grant access to your organisation; the client token is the only credential meant for browsers.

Where to go next

  1. Build the projects on a real account. Datadog offers a free trial; use a small test environment so you can experiment safely and keep an eye on usage.
  2. Pick a platform to go deeper: Kubernetes monitoring, serverless (Lambda), database monitoring, network monitoring or CI visibility.
  3. Learn the surrounding skills: SLO practice from site reliability engineering, structured logging, OpenTelemetry (Datadog can ingest OpenTelemetry data through the Agent or Collector), and Terraform.
  4. Study the official learning resources, such as the Datadog documentation, Learning Center and fundamentals courses. Certification exams exist; check the current exam guide, because topics and format change.
  5. Practise incident response with your team: run game days where someone breaks staging and the rest investigate.
flowchart LR
  A["Basics: Agent, tags"] --> B["Signals: metrics, logs, traces"]
  B --> C["Actions: monitors, SLOs, incidents"]
  C --> D["Scale: cost, code, security"]

A path from first install to running Datadog as a team.

The best Datadog skill is asking a good question of your data: "what changed, for whom, and since when?". Tags, traces and logs exist to answer exactly that.

Summary

  • Use the cheat sheet to connect problems with features.
  • Reading and writing queries, thresholds, burn rates and cardinality maths are the practical core skills.
  • Keep learning by building real projects, reading the official docs and running drills with your team.

Next: Back to the Datadog track