With agentic mode enabled, Dynatrace Assist personalization allows you to tailor how Dynatrace Assist responds to your prompts by providing additional context that's applied to every conversation. The additional context includes:
Your Profile and Custom Instructions fields are user-scoped, which means that each user can configure their personalization settings independently from their organization.
You can access the personalization feature only if you have agentic AI enabled in your environment and have sufficient permissions to use it.
To access the personalization feature
I'm a senior developer working with Infrastructure observability.Reply in German. Don't use emojis.The following limits apply to the Dynatrace Assist personalization feature:
Dynatrace Assist applies content filters and guardrails to all chat interactions. Since your profile information and your custom instructions are applied to every conversation, they might trigger our content filters or guardrails. This includes, but is not limited to attempts to:
If your personalized response is getting blocked, you can temporarily disable the personalization feature to test if your profile information or your custom instructions are causing this. If that's the case, we recommend that you rephrase the input for both fields and focus on your role, technical context, and response preferences.
The following types of information can increase Dynatrace Assist response quality and relevance:
I have no experience with Dynatrace or I have been using Dynatrace for 2 years and have a grasp on basic concepts. This information allows you to control whether Dynatrace Assist explains foundational concepts or opts to only explain advanced techniques and terminology.If you don't know what information you should provide when describing your profile to tailor your experience, you can use this list as a reference.
The following types of instructions allow Dynatrace Assist to adjust its response and style, and have measurable effects on response usefulness:
If you don't know what kind of instructions can help to tailor your experience, you can use this list as a reference.
Your Profile and Custom Instructions fields are stored as plain text and aren't subject to PII masking. Don't include any confidential credentials or sensitive personal data, such as:
You are responsible for ensuring that the content you enter in these fields is appropriate and doesn't contain information that should remain private. We strongly recommend limiting your profile to professional role, technical context, and response preferences only.
The following examples illustrate how different personas might configure their personalization settings.
Your Profile
I'm a Senior SRE at a financial services company. I manage Kubernetes clusters on AWS and use Dynatrace for full-stack observability. I'm experienced with DQL, Davis AI, and OpenTelemetry. My team is responsible for maintaining SLOs across 40+ microservices.
Custom Instructions
Be concise and technical – skip basic explanations. Always include DQL examples when relevant. Use metric keys and entity selectors in your answers rather than describing UI navigation. Respond in English.
Your Profile
I'm a Platform Engineer building an internal developer platform. I work primarily with Dynatrace workflows, Site Reliability Guardian, and OpenPipeline. My audience is development teams who are not Dynatrace experts. I have intermediate DQL skills and deep Kubernetes knowledge.
Custom Instructions
Explain concepts in a way I can relay to developers unfamiliar with Dynatrace. Prefer step-by-step instructions with code snippets. Include code or query examples when helpful. When suggesting solutions, prioritize Dynatrace-native approaches over third-party tools. Keep responses short unless I ask for detail. Avoid marketing language.
Your Profile
I'm a DevOps Engineer at a mid-size e-commerce company. I'm new to Dynatrace – we just started our rollout and I'm learning the platform. I have experience with Grafana, Prometheus, and AWS CloudWatch but I've never used DQL or Davis AI. I manage a mix of EC2 instances and EKS clusters running about 15 Java and Node.js services.
Custom Instructions
I'm still learning Dynatrace, so please explain platform-specific concepts and terminology when you use them. Link to relevant Dynatrace documentation when possible. When I ask about something I used to do in Grafana or Prometheus, show me the Dynatrace equivalent. Use step-by-step instructions with UI navigation paths rather than just DQL queries.