dynatrace/dynatrace-for-ai
12 skills8.2k installs1k starsGitHub
Install
npx skills add https://github.com/dynatrace/dynatrace-for-aiSkills in this repo
1Dt Dql Essentialsdt-dql-essentials is a compact command reference for Dynatrace Query Language (DQL), aimed at solo builders and small teams who already run workloads on Dynatrace and need correct pipeline syntax without leaving the agent. Install it when you are writing Grail queries, debugging slow services, correlating logs with metrics, or exploring dependency graphs via Smartscape. The skill walks through append, fetch, fields*, filter, join, limit, parse, summarize, timeseries, and related verbs with argument shapes so agents generate valid pipelines instead of guessing API names. It fits the Operate and Grow phases whenever analytics or incident response depends on ad-hoc DQL, and pairs naturally with Dynatrace AI skills that assume you can express filters and aggregations yourself. Intermediate complexity: you should know what a record pipeline is, but you do not need to memorize every optional parameter.745installs2Dt App Dashboardsdt-app-dashboards is an agent skill for builders who run workloads on Dynatrace and want dashboard-as-code style artifacts instead of clicking through the UI for every change. It encodes how to structure Example-style dashboard JSON: metadata name and type, content version, optional variables, and a tiles object mixing markdown documentation tiles with data tiles driven by Dynatrace Query Language. Data tiles can title metrics such as response time, attach lineChart visualizations, tune chartSettings including xAxisScaling to the analyzed timeframe, and define thresholds with enabled rules, comparators, labels, and CSS variable-based status colors. Units overrides let you normalize time-series fields for readable axes. Solo operators shipping APIs or agents with Dynatrace observability can use the skill to bootstrap SRE-facing boards that align with Dynatrace for AI workflows. Use it when you need to extend or duplicate service health views, wire WARN/CRITICAL cutoffs, or teach an agent the exact JSON shape Dynatrace expects for app dashboards.720installs3Dt Obs Logsdt-obs-logs is an agent skill for solo builders and small teams who already run workloads on Dynatrace and need fast, repeatable log investigations without hand-writing DQL from scratch every incident. It covers fetching logs, filtering by severity and entity, searching messages, computing error rates, and spotting patterns or trends—exactly the triggers users phrase as “show error logs,” “logs from last hour,” or “top error messages.” The skill explicitly scopes away from distributed tracing and span analysis, pointing those flows to dt-obs-tracing instead, and flags when Smartscape topology joins are required for host or cloud attributes. For indie operators wearing the on-call hat, it turns vague chat requests into structured observability queries suited to troubleshooting and ongoing monitoring rather than product docs or query explanation tasks.708installs4Dt Obs Problemsdt-obs-problems is an agent skill for Dynatrace DAVIS problem analysis aimed at solo builders and small teams running services in Dynatrace-managed environments. Install it when you need to answer operational questions in natural language—what is broken right now, what caused problem P-12345, how many users are affected, and whether an issue is recurring—without manually clicking through problem tiles in the UI. The skill organizes work into three workflows: active problem triage for open availability and performance incidents, root cause investigation that ties a display ID to root-cause and affected entities, and problem trending for history and blast-radius context. It is meant for telemetry-backed incident response where Dynatrace has already detected and correlated a problem, not for teaching product concepts, rewriting DQL from scratch, or replacing dedicated log or trace skills. Use it during on-call, post-deploy checks, and customer-impact assessments when you already have Dynatrace access and want structured, repeatable problem queries through your agent.697installs5Dt Obs Servicesdt-obs-services is a Dynatrace-focused agent skill that ships ready-made DQL patterns for .NET Common Language Runtime observability. Solo builders running .NET APIs or worker services on monitored hosts can paste and adapt queries for generation-level memory, GC invocation rates, and thread-pool queue depth without rebuilding metric catalogs from scratch. Each snippet maps to concrete dt.runtime.clr.* signals—memory.consumption, gc.collection_count, threadpool threads and queued items—scoped by process and process group. The skill fits operators who already ingest Dynatrace OneAgent data and want faster RCA when Gen2 pressure, JIT-related churn, or thread-pool starvation shows up in incidents. It is reference material, not an installer: you still need Dynatrace access and valid entity topology in smartscape.692installs6Dt Obs Tracingdt-obs-tracing is a Dynatrace-focused agent skill that teaches solo builders and small teams how to read database work inside distributed traces using DQL. Database calls show up as client spans with standard db.* attributes, and Dynatrace can aggregate many round-trips into a single span—so raw span counts understate load unless you apply sampling and aggregation multipliers. The skill walks through listing operations, mapping which services hit which db.system and namespace, and ranking hotspots by duration and extrapolated call volume. It fits indie operators who already instrument services with OpenTelemetry or Dynatrace OneAgent and need repeatable queries during performance investigations or cost-of-query reviews, without replacing full APM onboarding or schema design work.691installs7Dt App Notebooksdt-app-notebooks is an agent skill that supplies a Dynatrace app notebook definition agents can adapt when you need quick operational views of logs and host CPU. Solo builders running services on Dynatrace-gridded infrastructure can paste or extend the JSON to get table and chart sections without memorizing DQL field names or visualization state blocks. The template sets sensible defaults for timeframe, result caps, and Davis options so you can drill from raw logs into aggregated counts by severity and host. Use it after deploy when incidents or steady-state reviews need repeatable notebooks rather than one-off queries in chat. It does not replace full SRE runbooks or alert policies—extend sections and wire alerting separately.688installs8Dt Obs Hostsdt-obs-hosts is a Dynatrace-focused reference skill for solo builders and small teams running Kubernetes in production who need fast container inventory without writing DQL from scratch. It documents Smartscape queries that list containers with Kubernetes metadata, roll up distribution by cluster and namespace, and break counts down by workload kind such as deployments and daemonsets. The skill matters because mis-identifying workloads or assuming image tags exist in Smartscape leads to dead-end queries; the reference calls out that limitation upfront. Use it when you are operating live clusters and need lifecycle-oriented visibility—what is running where—not when you are still choosing a hosting story. It pairs with broader Dynatrace observability skills in the same repo and assumes you already ingest K8s into Dynatrace. Outputs are copy-paste DQL snippets and interpretation notes, not automated installs or dashboards.670installs9Dt Obs Kubernetesdt-obs-kubernetes is a reference skill that teaches an agent how to interrogate Kubernetes cluster and namespace inventory using Dynatrace-oriented query patterns. Solo builders running SaaS on K8s often lose time guessing which cluster, namespace, or workload actually hosts a regression; this skill catalogs list-and-count recipes for clusters, Kubernetes versions, nodes per cluster, namespaces (including system vs application), pod and workload distribution, and cross-cluster comparisons. It frames cluster and namespace entities as the foundation for deeper observability workflows rather than jumping straight to logs. Use it when you need topology maps, empty-namespace cleanup candidates, or to locate a named workload across a fleet. It pairs with Dynatrace MCP or similar observability backends referenced in the parent plugin. Complexity is intermediate because you must already ship on Kubernetes and have entity data ingested.660installs10Dt Obs AwsAWS Capacity Planning is a Dynatrace-oriented agent skill that gives solo operators ready DQL against Smartscape AWS entities. It organizes capacity work into compute, network, containers and serverless, databases and storage, and broader infrastructure—so you can see instance mix, ASG slack, and subnet IP availability before outages or costly over-provisioning. Invoke it when you run production on AWS and already ingest resources into Dynatrace, not when you are greenfield coding features. The skill is integration-shaped reference prose: paste queries, adapt fields, and iterate in Notebooks or dashboards. It suits Claude Code or Cursor users who manage their own observability stack and need repeatable capacity questions during Operate monitoring workflows.651installs11Dt Obs Frontendsdt-obs-frontends equips solo builders running SaaS or mobile web frontends in Dynatrace to go past default dashboards with DQL tuned for real-user monitoring. It supports CDN optimization by ranking countries where average duration stays high despite sufficient request volume, surfaces apps where elevated error rates coincide with slow requests, and frames regression checks after deployments—including contrast with synthetic monitoring called out in the skill’s use cases. Each query uses timeseries aggregations on dt.frontend.request.count, dt.frontend.request.duration, and dt.frontend.error.count, with fieldsAdd for human-readable seconds and percentages. Builders copy snippets into Notebooks or the query editor, adjust frontend.name filters, and shorten or lengthen the lookback window. Intermediate complexity assumes you already ingest browser RUM and know frontend entity naming. Use during Operate when alerts spike or after Ship when validating a release; Growth teams can reuse the same queries for lifecycle dashboards without relearning metric names.649installs12Dt Migrationdt-migration is a reference-oriented agent skill for builders running workloads on Dynatrace who need to migrate or reconcile auto-tagging keys with the semantic dictionary used by Dynatrace for AI and unified observability. It explains how each ManagementZoneDataSourceMeAttribute from the auto-tagging catalog should land on mass-data fields or Smartscape node attributes, including cases where no direct equivalent exists. The readme calls out tag-on-tag rule conditions, imported cloud and Kubernetes labels, and primary tag promotion (for example a selected label becoming primary_tags.team). Solo operators touching migration playbooks, management-zone cleanups, or AI-ready attribute consistency use it while operating production—not while sketching a prototype. Treat it as a lookup companion when renaming rules, auditing HOST_NAME versus detected names, or tracing which upstream tag source ultimately drives a rule.603installs