Course: Dynatrace SaaS · Module 1/8 — the Dynatrace Grail platform · Duration: ~45 min · Level: Beginner
Objectives:
- Place Grail at the centre of the Dynatrace SaaS platform and understand what it stores.
- Tell apart the main data types: logs, metrics, spans, events, entities.
- Find your way around the apps and know where to look for a feature.
- Understand the notions of environment, account and permissions.
Prerequisites: none, apart from access to a Dynatrace SaaS environment (the free trial is enough).
Before deploying anything, you need a map of the terrain. This first module introduces the Dynatrace Grail platform and the overall layout of Dynatrace SaaS: where data goes, how it is organised, which apps use it and who is allowed to see it. These ideas come back in every following module of the Dynatrace SaaS training course.
If you come from another observability tool, keep one guiding idea in mind: in the third-generation platform, there is no longer one store per data type, each with its own screens. Everything converges into a single store, queried with a single language.
What is the Dynatrace Grail platform?
Grail is the data store of Dynatrace SaaS. It receives data collected by OneAgent, cloud integrations, ingest APIs or OpenTelemetry, and keeps it without imposing a rigid schema up front. It is often called a “data lakehouse”: the flexibility of a data lake, with the ability to query it directly in a structured way.
Three practical consequences follow:
- A single query language. Logs, traces, metrics, events and entities are all queried with DQL (Dynatrace Query Language), which you will discover in module 3.
- Context is preserved. Records carry common attributes (host, service, Kubernetes namespace, trace) that let you move from one data type to another without copying and pasting identifiers.
- The schema is decided at read time. A field missing at ingestion can be extracted later, at query time, for example by parsing the content of a log line.
Buckets and retention
Inside Grail, records are stored in buckets. A bucket groups data of the same type and defines, among other things, its retention period. Default buckets exist for each type; an organisation can create others, for example to keep audit logs longer or to separate the data of two teams. Buckets are usually chosen at ingestion time, with OpenPipeline (module 6).
What types of data does Grail hold?
The table below summarises the main families you will work with during the course. The names in the right-hand column are the ones used in DQL queries.
| Family | What it describes | Example DQL source |
|---|---|---|
| Logs | Application and system log lines | logs |
| Traces | Spans of distributed requests | spans |
| Events | State changes, deployments, Davis problems | events, dt.davis.problems |
| Business events | Customer journey steps, transactions | bizevents |
| Metrics | Time series (CPU, latency, counters) | timeseries command |
| Entities | Hosts, services, processes, Kubernetes workloads | dt.entity.host, dt.entity.service… |
Entities deserve special attention. Dynatrace automatically builds a dependency model, historically called Smartscape: this process runs on that host, this service calls that other one. This model is what lets Davis link an application slowdown to infrastructure saturation (module 5).
How are the apps organised?
The SaaS interface is built as a set of apps, opened from the launcher or from search. Some are specialised by domain, others are cross-cutting tools. Remember three categories:
- Domain apps: infrastructure and hosts, Kubernetes, services, logs, user experience, problems. They offer ready-to-use views.
- Analysis tools: Notebooks to investigate, Dashboards to share (module 4).
- Platform tools: Workflows to automate (module 7), OpenPipeline ingestion settings (module 6), and the Hub to discover and install apps and extensions.
These apps run on AppEngine, a runtime that also lets Dynatrace, partners or your own team publish apps. Apps evolve independently: look for a feature by its name rather than by a menu path.
Davis, the built-in AI
Davis refers to the platform’s AI capabilities: anomaly detection, root cause analysis and more recent assistance features. It relies on Grail and on the entity model to produce the problems covered in module 5.
Environment, account and permissions
An environment is the space where data arrives and is viewed; it has its own web address. An account groups one or more environments, the users, the groups and the licence. Access management happens at account level, with groups to which you attach policies describing what is allowed: reading certain data, changing settings, running workflows.
This model allows fine-grained control, for example restricting a team to certain log buckets. In a company, “I can’t see the data” often comes from a policy, not from a collection problem. The official documentation details the policy syntax; we will not reproduce it here.
If you know the older Managed access model, described in users, groups and permissions in Managed, note that the two approaches differ: do not map roles across one to one.
Hands-on exercise
Step 1 — Open the trial environment
Create or open your Dynatrace SaaS trial environment and note its address. Find the home page, the app launcher and the search bar.
Expected result: you can find an app by its name within a few seconds.
Step 2 — Take an inventory of the apps
Browse the launcher and sort the apps you see into a personal table, using the module’s three categories (domain, analysis, platform). Open the Hub and look at the details page of two apps or extensions.
Expected result: a list of about ten classified apps, which you will add to as the course goes on.
Step 3 — Run a first query
Open the Notebooks app, create an empty notebook, add a query section and run:
fetch logs
| limit 10
On a new environment, the result may be empty: that is normal, no data has been collected yet. Then try fetch dt.davis.problems | limit 10.
Expected result: the query runs without error, with or without results. You will see these tables fill up after module 2.
Step 4 — Look at the buckets
Look in the environment settings for storage or bucket management, and review the list of default buckets and their retention.
Expected result: you identify at least one default bucket for logs and its retention period.
Quiz
- What is the role of Grail? A) An agent installed on hosts B) The platform’s single data store C) A dashboard tool D) A billing module
- What is a Grail bucket? A) A group of users B) A container for data of the same type, with its retention C) A type of dashboard D) A version of OneAgent
- Which language lets you query logs, spans and metrics? A) SQL B) PromQL C) DQL D) DPL
- A colleague cannot see certain logs even though they are ingested. What is the most likely cause? A) An access policy B) A failed OneAgent C) Retention that is too long D) A deleted dashboard
Answers
- B. Grail stores all observed data and makes it queryable with a single language.
- B. A bucket groups records of the same type and sets, among other things, how long they are kept.
- C. DQL queries every data type in Grail; DPL is used to describe parsing patterns.
- A. If the data is ingested, a policy restricting access to certain buckets or records is the most common cause.
Further reading
- Session Replay on Grail: replaying what the user experienced
- Log pattern analysis: from raw volume to DQL queries
- Understanding the Dynatrace Managed architecture, to compare with the older model
- Official documentation: docs.dynatrace.com, Grail and platform sections
← Course introduction · Course overview · Module 2: Deploying OneAgent and Dynatrace Operator →
Frequently asked questions
Does Grail completely replace Dynatrace’s older storage?
In the third-generation SaaS platform, Grail is the central store for logs, events, traces, metrics and entities. Depending on the age of your environment, some historical views may coexist; the documentation states the migration status for each data type.
Do I have to choose a bucket for every ingestion?
No. Without a specific rule, data goes into the default buckets. Creating dedicated buckets becomes useful when you have different retention or access needs.
Are all the platform’s apps included in the licence?
Consumption depends on your account’s licensing model and on data usage. Check with your Dynatrace contact or in account management before enabling a feature at scale.
You now have the map: Grail at the centre, buckets to organise, DQL to query, apps to present and policies to control access. The next module fills this empty store with real data.
span.observer is an independent community blog, not affiliated with Dynatrace LLC. This course does not replace the official Dynatrace University training and certifications. Check features against the official documentation for your version.
2 comments