Dynatrace University is one of the better vendor learning platforms in observability, and I say that as someone who finds most of the category tedious. The free tier is genuinely free and genuinely substantial. You get self-paced e-learning, video content and hands-on environments without a sales conversation first, which puts it ahead of several competitors who treat their training catalogue as a lead capture form with videos attached. The teaching quality is better than I expected.
Observability training has a failure mode where the vendor teaches you their user interface and calls it education, and Dynatrace mostly avoids it. The material on distributed tracing, on how spans relate to services and to the entities in the topology model, and on what a baseline actually represents, is conceptual work that survives outside the product. If you learn how Dynatrace models an environment as a dependency graph rather than a pile of metrics, you have learned something about observability architecture in general, and that is worth having. Davis, the causation engine, gets serious treatment and deserves it.
The pitch is that it identifies root cause rather than showing you correlated graphs and leaving you to guess, and the training is reasonably honest about where that works and where it depends on your instrumentation being complete. That honesty matters, because the failure mode of every automatic root cause system is that it confidently blames the thing it can see rather than the thing that broke. Grail, the data lakehouse underneath the newer platform, is where the training has had to work hardest, because the query language and the storage model are a significant departure from what long-time Dynatrace users knew. The migration material is decent.
The gap, and it is a large one, is cost. Dynatrace is expensive, its consumption pricing model is genuinely difficult to predict, and the single most common complaint I hear from teams running it is that their bill moved in a direction nobody forecast. Training that taught you how ingest volume, retention settings and query patterns drive cost would be enormously valuable and would take about two hours to produce. It is not there in any depth, and the reason is structural.
The vendor is not going to teach you to buy less of its product. That is not a Dynatrace-specific failing, it is true of every consumption-priced platform with an academy, but it is worth naming every time. The certification economics annoy me. Exams run in the region of two hundred to two hundred and fifty dollars per attempt depending on level and which source you believe, and Dynatrace does not publish this clearly, which is how you end up with a swarm of low quality third-party blog posts competing to tell you the number.
I could not get a definitive figure from the vendor without going further into the funnel than I was willing to, and that is itself a finding. Meanwhile vouchers get distributed at Perform, through partners and through customer success, often enough that anyone paying list price is simply not asking around. Three point seven. Strong free content, opaque paid credentialing, and a conspicuous silence about the bill.