Dynatrace
for LuumenAI
Problems, root-cause detail, and metrics beside the session.
Dynatrace already knows what is wrong; LuumenAI puts that next to the terminal that fixes it. Ask what problems are open, pull the root-cause analysis and affected entities into the thread, chart the metrics that matter, and silence a scope for a bounded window while governed work runs.
The Dynatrace toolbox
26 tools: 24 read, 2 write. Reads answer instantly. Writes require approval by default. Everything is logged.
- ReadList open problemsActive problems with impact, affected entities, and root-cause detail.
- ReadQuery metricsTime-series for hosts, services, and processes into the thread.
- ReadFetch entity detailHost, process, or service properties and recent events.
- ReadRead logsFiltered log lines for the entity you are looking at.
- ReadAutoadaptive Threshold Analysis AgentDetects anomalies in time series data by learning a single adaptive threshold from the data distribution that separates outliers from normal observations.
- ReadHelp AgentAnswers questions about Dynatrace documentation, product features, and terminology (e.g., Grail, OneAgent, Smartscape, pipeline, SLO, SLA, OpenPipeline).
- ReadGrail Query AgentConverts a natural language description into a valid Dynatrace Query Language (DQL) data retrieval query.
- ReadData Analysis AgentExecutes a valid DQL query against Grail and returns the result set.
- ReadDQL Explanation AgentReturns a human-readable explanation of a given DQL (Dynatrace Query Language) query.
- ReadDocument AgentSearches all dashboards and notebooks in the Dynatrace platform by title substring match.
- ReadTroubleshooting AgentSearches pre-indexed dashboards and notebooks in the "Problems troubleshooting" collection to find relevant troubleshooting guides.
- ReadSecurity Posture AgentRetrieve all compliance findings generated by Dynatrace Security Posture Management.
- ReadRuntime Vulnerability AgentRetrieve all open (muted and non-muted) security vulnerabilities generated by Dynatrace Runtime Vulnerability Analytics.
- ReadSmartscape AgentLooks up Dynatrace entity IDs by entity type and name substring.
- ReadKubernetes AgentRetrieves events for Kubernetes clusters within a specified time range.
- ReadRoot Cause Details AgentRetrieves full details of a single problem by its display ID.
- ReadSecurity Event Details AgentRetrieve the latest security event using a specific finding or scan ID or title.
- ReadSecurity Summary AgentQuery security events — both findings (vulnerabilities, detections, compliance issues, and misconfigurations) and scan events — from external products ingested into Dynatrace, as well as…
- ReadLog Pattern AgentAnalyze log lines coming from grail queries to find common patterns The logQuery input should always be a DQL query that fetches log lines from grail.
- ReadRoot Cause AgentQueries problems and returns up to 200 of the most recent results, filtered by status and time range.
- ReadSeasonal Baseline AgentDetects anomalies in time series data by learning normal behavior from historical data and establishing dynamic baselines that account for seasonal patterns (e.g., daily and weekly traffic cycles).
- ReadStatic Threshold Analysis AgentDetects anomalies in time series data by checking whether metric values violate a fixed threshold.
- ReadForecasting AgentPredicts future time series values using a statistical forecasting model based on historical data from a DQL query.
- ReadChangepoint AgentDetects changepoints in time series data, including outliers (spikes), level shifts, and significant trend changes, by computing novelty scores.
- WriteComment on problemAttach findings and actions to the Dynatrace problem. Approval by default
- WriteCreate maintenance windowSuppress alerting on a scope for a bounded period. Approval by default
One prompt, start to finish
What a governed Dynatrace run looks like inside Luumen.
Questions
Authorize once with API token. Luumen lists the scopes each action needs before you approve the connection, and credentials never appear in the chat.
Read actions answer immediately. Anything that writes — comment on problem, create maintenance window — is shown as a plan and requires approval by default. Administrators configure that per tool, so you decide exactly which actions can ever run unattended.
Dynatrace is available on the enterprise plan. Talk to us about enabling it for your account — the rest of the catalogue is available on every plan.
You decide. Actions are granted per agent, skill, and team, and per environment — production is not staging. Read access can be broad while writes stay narrow.
Every call to Dynatrace — read or write, approved or declined — is recorded with the actor, the input, and the result, and can be linked to the ticket or change record.
Often connected alongside
All integrations
Put Dynatrace to work with Luumen
Connect in minutes. Every action scoped, approved, and audited from day one.