Create and manage Datadog dashboards as code using the datadog_dashboard Terraform resource

Steps

1. Configure the DataDog/datadog Terraform provider with an API key and application key (for example via DD_API_KEY and DD_APP_KEY environment variables, or api_key/app_key provider arguments) rather than hardcoding credentials in .tf files.
2. Define a datadog_dashboard resource with the two required top-level arguments: title (String) and layout_type (String, either "ordered" or "free"); add an optional description and tags (only team:<name> tags are supported).
3. Add one or more widget blocks inside the resource, each containing exactly one nested *_definition block for the widget type you want (e.g. timeseries_definition, query_value_definition, alert_graph_definition, change_definition), with widget-specific required fields such as request/query blocks and a title.
4. For ordered dashboards, optionally set reflow_type to "auto" (widgets have no explicit layout) or "fixed" (each widget must define its own layout) — reflow_type only applies when layout_type = "ordered".
5. Use restricted_roles (a set of role UUIDs) to control who can edit the dashboard; do not use is_read_only, which is deprecated and non-functional.
6. Run terraform plan and terraform apply; the resource's id becomes the Datadog dashboard ID, and dashboard_lists can optionally attach the dashboard to existing dashboard lists.