> ## Documentation Index > Fetch the complete documentation index at: https://ravion-b90c0359-devin-1788408180-cross-module-output-refs.mintlify.site/llms.txt > Use this file to discover all available pages before exploring further. # Get DeploymentResource by id > Retrieve a single deployed Terraform resource by ID. Returns its full configuration and current state. GET /deployment-resources/{id} in the Ravion API… ## OpenAPI ````yaml https://api.ravion.com/openapi.yaml get /deployment-resources/{id} openapi: 3.0.0 info: title: Ravion version: 0.0.0 servers: - url: https://api.ravion.com security: - BearerAuth: [] tags: - name: Projects - name: Environments - name: FeatureFlags - name: Pipelines - name: PipelineRuns - name: TerraformResources - name: TerraformExecutionSummaries - name: PipelineStepExecutions - name: AwsCloudWatch - name: PipelineVersions - name: Organizations - name: Stacks - name: StackWorkspaces - name: Auth - name: OAuth - name: User - name: Health - name: Memberships - name: ServiceAccounts - name: AwsDefaultNetworks - name: AwsAccounts - name: ApiKeys - name: AwsAmp - name: EksPrometheus - name: EksLoki - name: ExecutionEnvironments - name: ModuleDefinitions - name: ModuleCategories - name: ModuleVersions - name: ModuleInstances - name: DefaultValueDefinitions - name: DefaultValues - name: CodeSources - name: Github - name: Gitlab - name: Git - name: Values - name: Deployments - name: DeploymentResources - name: InfrastructureEvents - name: WebSocket - name: Domains - name: AcmCertificates - name: Describe - name: Reports - name: BillingPlans - name: BillingAccounts - name: BillingCycles - name: BillingUsageSummaries - name: BillingAddOns - name: BillingInvoices paths: /deployment-resources/{id}: get: tags: - DeploymentResources summary: Get DeploymentResource by id description: >- Retrieve a single deployed Terraform resource by ID. Returns its full configuration and current state. operationId: GetDeploymentResource parameters: - in: path name: id required: true schema: type: string responses: '200': content: application/json: schema: additionalProperties: false properties: data: $ref: '#/components/schemas/DeploymentResource.DetailResponse' required: - data type: object description: The request has succeeded. '401': content: application/json: schema: $ref: '#/components/schemas/Errors.UserFacingErrorData' description: You are not authenticated '403': content: application/json: schema: $ref: '#/components/schemas/Errors.UserFacingErrorData' description: You do not have permission to access this resource. '404': content: application/json: schema: $ref: '#/components/schemas/Errors.UserFacingErrorData' description: The server cannot find the requested resource. '500': content: application/json: schema: $ref: '#/components/schemas/Errors.UserFacingErrorData' description: Server error components: schemas: DeploymentResource.DetailResponse: additionalProperties: false properties: data: $ref: '#/components/schemas/DeploymentResource.Detail' required: - data type: object Errors.UserFacingErrorData: additionalProperties: false description: |- User-facing error presentation data. This is what the API returns to the frontend after formatting ErrorData using CEL templates from the error registry. Used for both: - Error fields on domain models (e.g., PipelineRun.error) - API error response bodies (HTTP 4xx/5xx responses) properties: action: allOf: - $ref: '#/components/schemas/Errors.Action' description: Optional action to help resolve the error code: description: Full error code, e.g., "Ravion:Pipeline:NOT_FOUND" type: string description: description: Additional description with more details type: string details: description: Structured details rendered as user-facing sections. items: $ref: '#/components/schemas/Errors.UserFacingErrorDetailSection' type: array isInternal: description: >- Indicates whether this error is internal (only set when ShowInternal=true). This allows SUPERADMINs to identify internal errors while viewing full details. type: boolean message: description: Main error message (required) type: string metadata: additionalProperties: {} description: >- Error params/metadata. Stripped for internal errors unless superadmin. type: object requestId: description: Request ID for correlating errors with server logs. type: string required: - code - message type: object DeploymentResource.Detail: additionalProperties: false description: Detail shape with ancillary fields the drawer needs but lists don't. properties: awsAccountId: description: |- Database ID of the AWS account that owns the cluster / function / distribution. type: string data: $ref: '#/components/schemas/DeploymentResourceData' desiredStatus: type: string ec2Instance: allOf: - $ref: '#/components/schemas/DeploymentResource.Ec2InstanceDetail' description: |- Live configuration and health for EC2_INSTANCE. Absent for other resource kinds and when the AWS account can no longer be resolved. error: $ref: '#/components/schemas/Errors.UserFacingErrorData' externalId: description: |- AWS identity for this resource. Task ARN for ECS tasks, request ID for Lambda invocations (future), distribution ID for CloudFront distributions under a static deploy (future). type: string firstSeenAt: $ref: '#/components/schemas/UtcDateTime' id: type: string lastEventAt: $ref: '#/components/schemas/UtcDateTime' lastStatus: description: >- Generic status string. Kind-specific ladders: ECS_TASK → PROVISIONING | PENDING | RUNNING | DEPROVISIONING | STOPPED EC2_INSTANCE → PENDING | IN_SERVICE | TERMINATING | TERMINATED | DETACHING | DETACHED | STANDBY LAMBDA_INVOCATION (future) → INIT | RUNNING | SUCCESS | ERROR | TIMEOUT STATIC_CLOUDFRONT_DISTRIBUTION (future) → PENDING_UPDATE | IN_PROGRESS | DEPLOYED | FAILED type: string lifecycleType: allOf: - $ref: '#/components/schemas/Deployments.DeploymentResourceLifecycleType' description: |- Lifecycle type within the parent deployment — RUNTIME for the long-lived service-managed task, PRE_DEPLOY / POST_DEPLOY for the optional hook tasks. logStreams: description: |- Per-container CloudWatch log streams for ECS_TASK. Empty / absent for kinds without per-container logs. items: $ref: '#/components/schemas/DeploymentResource.EcsTaskLogStream' type: array moduleDeploymentId: description: |- Parent ModuleDeployment id — the deploy that spawned this resource. type: string moduleInstanceId: description: |- Optional linkage to the module instance (denormalized from the deploy so same-resource-across-deploys queries are cheap). type: string startedAt: $ref: '#/components/schemas/UtcDateTime' stoppedAt: $ref: '#/components/schemas/UtcDateTime' type: allOf: - $ref: '#/components/schemas/Deployments.DeploymentResourceType' description: Runtime kind of the resource — drives the shape of `data`. required: - id - type - lifecycleType - moduleDeploymentId - externalId - firstSeenAt - lastEventAt - data type: object Errors.Action: additionalProperties: false description: Action to help user resolve the error properties: label: description: Button/link label text type: string url: description: URL to navigate to for resolution type: string required: - label - url type: object Errors.UserFacingErrorDetailSection: additionalProperties: false description: Structured user-facing error detail section. properties: items: description: List of detail values for this section. items: type: string type: array object: additionalProperties: {} description: Structured detail payload for object rendering. type: object render: description: Rendering hint for clients. Valid values are list or object. type: string title: description: Detail section title shown in the UI. type: string required: - title - render type: object DeploymentResourceData: description: |- Discriminated union over `data` payloads for the resource envelope. Variants live in `@fc/schemas/deployments/tsp/aws-ecs.tsp`; the union here only wires them up with the shared `type` discriminator so JSON decodes unambiguously. Adding a new deployment kind adds one variant here + one enum value on `Type`. discriminator: mapping: EC2_INSTANCE: $ref: '#/components/schemas/Deployments.AwsEc2InstanceData' ECS_TASK: $ref: '#/components/schemas/Deployments.EcsTaskData' K8S_WORKLOAD: $ref: '#/components/schemas/Deployments.EksWorkloadResourceData' propertyName: type oneOf: - $ref: '#/components/schemas/Deployments.EcsTaskData' - $ref: '#/components/schemas/Deployments.AwsEc2InstanceData' - $ref: '#/components/schemas/Deployments.EksWorkloadResourceData' type: object DeploymentResource.Ec2InstanceDetail: additionalProperties: false description: |- On-demand AWS projection used by the EC2 instance drawer. The persisted DeploymentResource remains the source of truth for membership and release state; this payload supplies volatile configuration and health. properties: architecture: type: string autoScaling: $ref: '#/components/schemas/DeploymentResource.Ec2AutoScalingDetail' availabilityZone: type: string capacity: $ref: '#/components/schemas/DeploymentResource.Ec2InstanceTypeDetail' iamInstanceProfileArn: type: string imageId: type: string instanceType: type: string launchTime: $ref: '#/components/schemas/UtcDateTime' listeners: items: $ref: '#/components/schemas/DeploymentResource.Ec2ListenerDetail' type: array loadBalancers: items: $ref: '#/components/schemas/DeploymentResource.Ec2LoadBalancerDetail' type: array monitoringState: type: string name: type: string platformDetails: type: string privateDnsName: type: string privateIpAddress: type: string publicDnsName: type: string publicIpAddress: type: string securityGroups: items: $ref: '#/components/schemas/DeploymentResource.Ec2SecurityGroup' type: array ssm: $ref: '#/components/schemas/DeploymentResource.Ec2SsmDetail' state: type: string stateTransitionReason: type: string status: $ref: '#/components/schemas/DeploymentResource.Ec2StatusDetail' subnetId: type: string tags: items: $ref: '#/components/schemas/DeploymentResource.Ec2Tag' type: array targetGroup: $ref: '#/components/schemas/DeploymentResource.Ec2TargetGroupDetail' targetHealth: $ref: '#/components/schemas/DeploymentResource.Ec2TargetHealthDetail' unavailableSections: description: |- AWS sources that could not be read. Other populated sections remain valid and render normally. items: type: string type: array vpcId: type: string required: - securityGroups - tags - loadBalancers - listeners - unavailableSections type: object UtcDateTime: description: UTC RFC3339 timestamp with a required Z suffix format: date-time pattern: ^\d{4}-\d{2}-\d{2}T\d{2}:\d{2}:\d{2}(\.\d{1,9})?Z$ type: string x-go-type: apitypes.UTCDateTime x-go-type-import: name: apitypes path: ravion/api-go/pkg/apitypes Deployments.DeploymentResourceLifecycleType: description: |- Which lifecycle role a DeploymentResource plays within its ModuleDeployment. RUNTIME is the long-lived service-managed task; PRE_DEPLOY / POST_DEPLOY are the optional hook tasks. enum: - RUNTIME - PRE_DEPLOY - POST_DEPLOY type: string DeploymentResource.EcsTaskLogStream: additionalProperties: false properties: containerName: type: string logGroupName: type: string logStreamPrefix: type: string region: type: string streamName: type: string required: - containerName - logGroupName - region - logStreamPrefix - streamName type: object Deployments.DeploymentResourceType: description: |- Which runtime type a DeploymentResource row represents. Drives the shape of its `data` JSON column. Mirrors the `type` discriminator on the JSON payload variants (EcsTaskData.type = "ECS_TASK", …) so the column and the JSON tag carry the same value. enum: - ECS_TASK - EC2_INSTANCE - K8S_WORKLOAD type: string Deployments.AwsEc2InstanceData: additionalProperties: false description: Current state for one EC2 instance belonging to the module's ASG. properties: autoScalingGroupName: type: string availabilityZone: type: string awsAccountNumber: type: string instanceArn: description: Full account- and region-qualified EC2 instance ARN. type: string instanceId: description: Human-readable EC2 instance id (`i-…`). type: string outputTail: description: >- Tail of the latest SSM command output, bounded by the activity before persistence. Failed releases use this for diagnostics; successful releases expose it in the instance drawer's deploy-output tab. type: string region: type: string releaseCommandId: description: |- SSM command that most recently attempted this release on this concrete instance. Catch-up commands for replacement instances have their own command id, distinct from the original deployment command. type: string releaseCompletedAt: format: date-time type: string releaseExecutionStartedAt: description: Earliest time an SSM document step began executing on the instance. format: date-time type: string releaseGate: allOf: - $ref: '#/components/schemas/Deployments.AwsEc2InstanceReleaseGate' description: >- Present only for instances launched behind the versioned Ravion release gate. releaseRequestedAt: description: Time Systems Manager sent the invocation to this managed instance. format: date-time type: string releaseResponseCode: description: >- Exit code retained from the invocation's failed or final document step. format: int32 type: integer releaseStartedAt: description: |- First time the current release entered an active SSM state on this instance. This is recorded when SendCommand is accepted. format: date-time type: string releaseStatus: allOf: - $ref: '#/components/schemas/Deployments.AwsEc2InstanceReleaseStatus' description: |- Release application state for the deployment referenced by the DeploymentResource envelope's `moduleDeploymentId`. releaseStatusDetails: description: >- Detailed AWS invocation status retained for actionable timeline copy. type: string subnetId: type: string type: enum: - EC2_INSTANCE type: string required: - type - instanceId - instanceArn - autoScalingGroupName - awsAccountNumber - region - releaseStatus type: object Deployments.EcsTaskData: additionalProperties: false description: |- Typed payload for DeploymentResource.data when type == ECS_TASK. Carries every ECS-specific field the old EcsTask table held as typed columns. The generic DeploymentResource envelope owns the fields merge needs to reason about across types (lastStatus, desiredStatus, startedAt, stoppedAt, firstSeenAt, lastEventAt, externalId) as real SQL columns. Everything else — ECS-specific descriptors and phase timestamps — lives in this JSON. properties: availabilityZone: type: string capacityProviderName: type: string clusterArn: type: string connectivity: type: string containerInstanceArn: type: string deprovisioningAt: format: date-time type: string exitCode: description: |- Exit code of the essential container that ended the task. Set once when the task reaches STOPPED with a non-null `exitCode` on any of its containers. format: int32 type: integer healthStatus: type: string launchType: type: string pendingAt: format: date-time type: string platformVersion: type: string provisioningAt: format: date-time type: string pullStartedAt: format: date-time type: string pullStoppedAt: format: date-time type: string runTaskOverrides: allOf: - $ref: '#/components/schemas/Module.EcsRunTaskOverrides' description: |- Snapshot of the AWS RunTask `TaskOverride` payload the deploy step passed when launching this task. Only populated for PRE_DEPLOY / POST_DEPLOY hook tasks — the long-lived RUNTIME service task inherits the task definition verbatim and never carries an override payload. Reuses `Module.EcsRunTaskOverrides` directly — the same model the user authors in their YAML under `pre_deploy:` / `post_deploy:` — so config and runtime snapshot share one type with no parallel struct and no drift when AWS adds a new override field. runningAt: format: date-time type: string startedBy: type: string stopCode: type: string stoppedReason: type: string targetGroupSnapshot: allOf: - $ref: '#/components/schemas/Deployments.EcsTaskTargetGroupSnapshot' description: |- Frozen target-group evidence captured at task-stop. The event ingester runs the selector once against a fresh AWS fetch, picks the single most diagnostic TG, and persists the result here. Absent while the task is running, for workers without a load balancer, or when the task stopped before any TG registration was observed. The read path is a pure unmarshal — no runtime selection. taskArn: type: string taskDefinitionArn: type: string taskDefinitionFamily: type: string taskDefinitionRevision: format: int32 type: integer taskGroup: type: string taskTargetId: description: |- Normalized target-group registration key for this task. awsvpc: private IPv4 extracted from the ENI attachment snapshot. EC2 bridge/host: the containerInstanceArn (ECS registers the instance in the target group rather than a per-task IP). Used by the drawer to filter target-health rows to exactly this task. type: string type: enum: - ECS_TASK type: string required: - type - taskArn - clusterArn - taskDefinitionArn type: object Deployments.EksWorkloadResourceData: additionalProperties: false description: |- Typed payload for `DeploymentResource.data` when type == K8S_WORKLOAD. The generic DeploymentResource envelope owns what the platform reasons about across kinds — `externalId` (the composite workload key), `lastStatus` (Operator's rollout classification), `moduleDeploymentId`, `moduleInstanceId`, `lastEventAt`. Everything Kubernetes-specific lives here. properties: availableReplicas: format: int32 type: integer currentReplicas: format: int32 type: integer desiredReplicas: format: int32 type: integer diagnoses: description: >- What Operator has concluded is wrong with this workload's pods, newest first, bounded — a workload that has been failing all week must not grow an unbounded payload. Unlike `regression` and `drift` this is a LIST rather than a single object, because a workload really can have several unrelated findings at once (an image that will not pull on one container while another is OOM-killed), and collapsing them would hide whichever one Ravion happened to rank lower. Each entry is guarded by its own timestamps rather than by a shared stream sequence — see the note on `EksWorkloadDiagnosis`. items: $ref: '#/components/schemas/Deployments.EksWorkloadDiagnosis' type: array drift: allOf: - $ref: '#/components/schemas/Deployments.EksWorkloadDrift' description: |- Current drift finding, absent while the agent has never reported on this workload's drift. Guarded by its own `drift.sequence`. everReady: description: |- False means no replica of the current generation has ever become ready — the difference between a deploy that never came up and one that came up and later degraded. type: boolean generation: format: int64 type: integer kind: description: Deployment / StatefulSet / DaemonSet / … type: string message: type: string name: type: string namespace: type: string observedGeneration: format: int64 type: integer readyReplicas: format: int32 type: integer reason: type: string regression: allOf: - $ref: '#/components/schemas/Deployments.EksWorkloadRegression' description: |- Current regression episode, absent while the workload has never regressed. Guarded by its own `regression.sequence`. releaseName: description: >- Helm attribution, verbatim from the cluster. These are what the agent reports; `moduleDeploymentId` on the envelope is what the control plane resolved them to. type: string releaseNamespace: type: string sequence: description: >- The agent's monotonic ordering key for the ROLLOUT stream, seeded from a nanosecond wall-clock reading so it keeps increasing across an agent restart. A report whose sequence is not strictly greater than the stored one is a silent no-op. format: int64 type: integer stalledForSeconds: format: int32 type: integer stuckReason: type: string type: enum: - K8S_WORKLOAD type: string updatedReplicas: format: int32 type: integer workloadUid: type: string required: - type - kind - namespace - name - releaseName - releaseNamespace - sequence type: object DeploymentResource.Ec2AutoScalingDetail: additionalProperties: false description: Auto Scaling's current view of the instance. properties: healthStatus: type: string launchTemplateId: type: string launchTemplateName: type: string launchTemplateVersion: type: string lifecycleState: type: string protectedFromScaleIn: type: boolean type: object DeploymentResource.Ec2InstanceTypeDetail: additionalProperties: false description: Static capacity characteristics of the concrete EC2 instance type. properties: architecture: type: string defaultVcpus: format: int32 type: integer memoryMib: format: int64 type: integer networkPerformance: type: string type: object DeploymentResource.Ec2ListenerDetail: additionalProperties: false description: Listener rule that forwards traffic to the selected target group. properties: hostHeaders: items: type: string type: array isDefaultRule: type: boolean listenerArn: type: string loadBalancerArn: type: string loadBalancerDnsName: type: string pathPatterns: items: type: string type: array port: format: int32 type: integer protocol: type: string rulePriority: format: int32 type: integer required: - listenerArn - loadBalancerArn - hostHeaders - pathPatterns - isDefaultRule type: object DeploymentResource.Ec2LoadBalancerDetail: additionalProperties: false description: Load balancer attached to the target group. properties: dnsName: type: string loadBalancerArn: type: string loadBalancerName: type: string scheme: type: string type: type: string required: - loadBalancerArn type: object DeploymentResource.Ec2SecurityGroup: additionalProperties: false description: One security group attached to the instance's primary network path. properties: groupId: type: string groupName: type: string required: - groupId type: object DeploymentResource.Ec2SsmDetail: additionalProperties: false description: Systems Manager's current view of the managed node. properties: agentVersion: type: string associationStatus: type: string computerName: type: string isLatestAgentVersion: type: boolean lastPingAt: $ref: '#/components/schemas/UtcDateTime' pingStatus: type: string platformName: type: string platformType: type: string platformVersion: type: string type: object DeploymentResource.Ec2StatusDetail: additionalProperties: false description: EC2 platform status checks and scheduled maintenance. properties: attachedEbsStatus: type: string events: items: $ref: '#/components/schemas/DeploymentResource.Ec2ScheduledEvent' type: array instanceStatus: type: string systemStatus: type: string required: - events type: object DeploymentResource.Ec2Tag: additionalProperties: false description: |- One EC2 tag. Kept as an array so duplicate/empty values preserve the exact AWS response and the UI can render stable key/value rows. properties: key: type: string value: type: string required: - key type: object DeploymentResource.Ec2TargetGroupDetail: additionalProperties: false description: Target-group configuration surrounding the selected instance target. properties: healthCheckPath: type: string healthCheckPort: type: string healthCheckProtocol: type: string loadBalancerArns: items: type: string type: array matcherGrpcCode: type: string matcherHttpCode: type: string port: format: int32 type: integer protocol: type: string protocolVersion: type: string targetGroupArn: type: string targetGroupName: type: string targetType: type: string required: - targetGroupArn - loadBalancerArns type: object DeploymentResource.Ec2TargetHealthDetail: additionalProperties: false description: |- Load-balancer target health for this instance, when the module has a target group. properties: description: type: string healthCheckPort: type: string port: format: int32 type: integer reason: type: string state: type: string targetGroupArn: type: string required: - targetGroupArn type: object Deployments.AwsEc2InstanceReleaseGate: additionalProperties: false description: Persisted state for a versioned Ravion Auto Scaling launch gate. properties: commandId: description: SSM command applying the selected release. type: string completedAt: format: date-time type: string eventId: description: EventBridge event that created this gate. type: string failureReason: type: string lifecycleHookName: description: Auto Scaling lifecycle hook used for heartbeat and completion calls. type: string releaseDeploymentId: description: >- Successful deployment selected after acquiring the deployment-manager lock. type: string startedAt: format: date-time type: string status: $ref: '#/components/schemas/Deployments.AwsEc2InstanceReleaseGateStatus' required: - eventId - lifecycleHookName - status type: object Deployments.AwsEc2InstanceReleaseStatus: description: Whether the current release has been applied to a live EC2 instance. enum: - UNKNOWN - PENDING - IN_PROGRESS - SUCCESS - FAILED type: string Module.EcsRunTaskOverrides: additionalProperties: false description: >- Full AWS ECS RunTask request payload — both the `TaskOverride` fields and the top-level RunTask parameters that aren't part of `TaskOverride` (launchType, networkConfiguration, …). Flattened into one model so users can express any RunTask shape in one YAML block. Used as: • the user-authored hook config in the module YAML (under `pre_deploy:` / `post_deploy:`), • the runtime snapshot persisted onto `DeploymentResource.data.runTaskOverrides` so the task drawer can show operators exactly what was sent to AWS. Deploy-managed RunTask fields (`cluster`, `taskDefinition`, `count`, `clientToken`) are intentionally not exposed — the deploy step controls them. https://docs.aws.amazon.com/AmazonECS/latest/APIReference/API_RunTask.html https://docs.aws.amazon.com/AmazonECS/latest/APIReference/API_TaskOverride.html properties: capacity_provider_strategy: description: |- Capacity-provider strategy — overrides `launch_type` when set. Inherits from the parent service when omitted. items: $ref: '#/components/schemas/Module.EcsCapacityProviderStrategyItem' type: array x-templateable: true container_overrides: description: |- Per-container overrides. Each entry's `name` matches a container in the task definition. AWS RunTask supports overriding any subset of containers; the array can be empty when only task-level overrides are needed. items: $ref: '#/components/schemas/Module.EcsContainerOverride' type: array x-templateable: true cpu: description: |- Task-level CPU override (string for AWS compatibility — same format as task-definition `cpu`, e.g. `"256"`, `"1024"`). type: string x-templateable: true enable_ecs_managed_tags: description: Whether to enable AWS-managed `aws:ecs:` tags on the task. type: boolean x-templateable: true enable_execute_command: description: >- Whether to enable ECS Exec on the task (for `aws ecs execute-command`). type: boolean x-templateable: true ephemeral_storage: allOf: - $ref: '#/components/schemas/Module.EcsEphemeralStorage' description: Fargate ephemeral storage size override (20–200 GiB). x-templateable: true execution_role_arn: description: |- IAM execution role ARN override — what ECS uses to pull the image, push logs, and read secrets at task launch. type: string x-templateable: true group: description: |- Logical group name — used for placement constraints (`memberOf(group:)`). Defaults to `family:` when AWS picks. type: string x-templateable: true launch_type: description: |- Launch type: FARGATE, EC2, or EXTERNAL. When omitted the hook inherits the parent service's launch type. Mutually exclusive with `capacity_provider_strategy` — set one or the other. enum: - FARGATE - EC2 - EXTERNAL type: string x-templateable: true memory: description: Task-level memory override (MiB, string format). type: string x-templateable: true network_configuration: allOf: - $ref: '#/components/schemas/Module.EcsNetworkConfiguration' description: |- VPC / subnet / security-group config. Required for `awsvpc` network-mode task definitions; inherited from the parent service when omitted. x-templateable: true placement_constraints: description: |- Placement constraints — limits which container instances can run the task. EC2/EXTERNAL launch types only. items: $ref: '#/components/schemas/Module.EcsPlacementConstraint' type: array x-templateable: true placement_strategy: description: |- Placement strategy — how tasks get distributed across instances. EC2/EXTERNAL launch types only. items: $ref: '#/components/schemas/Module.EcsPlacementStrategy' type: array x-templateable: true platform_version: description: |- Fargate platform version (e.g. `"1.4.0"`, `"LATEST"`). Inherits from the parent service when omitted. type: string x-templateable: true propagate_tags: description: How tags propagate from task definition / service onto the task. enum: - TASK_DEFINITION - SERVICE - NONE type: string x-templateable: true reference_id: description: |- Caller-supplied reference id (UUID-shaped). Echoed back on the resulting Task and useful for client-side correlation. type: string x-templateable: true tags: description: Metadata tags applied to the task. items: $ref: '#/components/schemas/Module.EcsTag' type: array x-templateable: true task_role_arn: description: |- IAM task role ARN override — what the container assumes when calling AWS APIs from inside the task. type: string x-templateable: true timeout: description: |- Wall-clock timeout in seconds before the task is stopped and the deployment fails. Only meaningful when used as a pre/post deploy hook config — ignored on the runtime snapshot. Default: 1800 (30 minutes). format: int32 minimum: 1 type: integer x-templateable: true volume_configurations: description: |- Per-task volume bindings — name must match a volume in the task definition. Used for Fargate-managed ephemeral EBS volumes. items: $ref: '#/components/schemas/Module.EcsTaskVolumeConfiguration' type: array x-templateable: true type: object Deployments.EcsTaskTargetGroupSnapshot: additionalProperties: false description: |- Frozen target-group evidence for a single task, computed once at task-stop by the selector and persisted on `EcsTaskData.targetGroupSnapshot`. Singular — the selector picks the single best TG (state tier → source priority → proximity) at write time. The read path is a plain unmarshal; no runtime ranking. Only persisted when the task's own target was observed in the TG; otherwise the drawer collapses the Targets column. properties: capturedAt: format: date-time type: string evidenceQuality: $ref: '#/components/schemas/Deployments.EcsEvidenceQuality' healthCheck: $ref: '#/components/schemas/Deployments.EcsTargetHealthCheckConfig' source: $ref: '#/components/schemas/Deployments.EcsTargetHealthSource' targetGroupArn: type: string targetGroupName: type: string taskTarget: allOf: - $ref: '#/components/schemas/Deployments.EcsTaskTargetObservation' description: |- This task's own target in the TG. Always populated — the selector only writes a snapshot when the task's target was observed. required: - targetGroupArn - taskTarget - capturedAt - source - evidenceQuality type: object Deployments.EksWorkloadDiagnosis: additionalProperties: false description: >- One thing Operator concluded is wrong with this workload, in Ravion's words alongside the cluster's own. IDENTITY IS `(code, reason)`, and the pod is deliberately not part of it. That is the agent's own deduplication key minus the workload, which is the row: three replicas failing identically is ONE problem, and keying on the pod would turn a rollout of a broken image into a wall of identical findings. `podName` records which pod the finding was last seen on, as evidence, not as identity. Both the raw Kubernetes facts (`reason`, `message`) and Ravion's reading of them (`cause`, `suggestedAction`) are kept. A user who suspects Ravion has misread their cluster must be able to check, and it is what makes an `UNKNOWN` code useful rather than embarrassing. ORDERED BY `observedAt`, NOT BY A SEQUENCE. The diagnosis stream carries no monotonic counter — findings are independent of one another, and one shared counter would let a crash loop's report order against an unrelated image-pull failure. Each finding is guarded by its own timestamps instead: a report older than what is stored for the same `(code, reason)` is a silent no-op, which is what stops a redelivered raise from resurrecting a finding the user was already told had cleared. properties: cause: description: >- What the failure means, in Ravion's language. Prose — nothing may parse it. type: string code: description: >- CRASH_LOOP_BACKOFF | IMAGE_PULL_FAILED | OOM_KILLED | … Free-form for the same version-skew reason every other agent-reported classification is: an agent newer than the control plane will report codes this schema has never heard of, and refusing them would fail closed on exactly the clusters running the newest code. type: string container: description: >- The container the finding concerns, absent when it is about the pod as a whole (a scheduling failure has no container). type: string crashLog: allOf: - $ref: '#/components/schemas/Deployments.EksWorkloadDiagnosisLog' description: >- What the container printed before it died, when Operator captured it. ABSENT means no capture was attempted — a scheduling failure has no container to read. A capture that was attempted and failed is present with `available: false`, which is a different statement and must stay one. details: additionalProperties: type: string description: >- The structured facts behind `cause` and `suggestedAction`, keyed by the agent's detail names (`exit_code`, `image`, `probe`, …). Every value is a string even when it is a number: which keys are present varies by code, and one map with one value type is far easier to render generically than a union. type: object firstObservedAt: description: >- When this occurrence was FIRST seen. A recurrence after a resolution is a new occurrence with a new `firstObservedAt`, so a finding standing since Tuesday does not read as something happening now. format: date-time type: string message: description: The Kubernetes message, verbatim. type: string observations: description: >- How many times the agent confirmed the condition still held, including the observation that raised it. format: int32 type: integer observedAt: description: |- When the finding was last confirmed. This is the value staleness is judged against. format: date-time type: string podName: description: >- The pod the finding was last observed on. Evidence, not identity — see the note above — and frequently a pod that no longer exists. type: string podNamespace: type: string reason: description: >- The Kubernetes reason, verbatim — never rewritten, never normalised. It is the thing a user will search for, and the second half of the finding's identity. type: string resolved: description: >- True once the agent reported the condition no longer holds. The finding is kept rather than deleted, for two reasons: a user who saw a banner deserves to see that it cleared, and the tombstone is what makes a redelivered raise a no-op instead of a resurrection. type: boolean resolvedAt: description: >- When the agent reported the condition had cleared. Set only alongside `resolved`. format: date-time type: string severity: description: >- error | warning. Two values on purpose — a third would invite arguments about which bucket a failure belongs in without changing what anybody does about it. Free-form all the same. type: string suggestedAction: description: |- What to do next. It names a specific thing to look at or change; restating the failure is not an action. Prose. type: string title: description: |- The agent's short human label for the code, so a UI has something to render for a code it does not recognise. type: string required: - code type: object Deployments.EksWorkloadDrift: additionalProperties: false description: |- The drift half of an EKS workload's live state: where the live object no longer matches the spec Ravion deployed. Report-only — Ravion never reverts drift, and the base agent RBAC has no write verb to do it with. `drifted: false` is NOT an absence of news. It is the clear that takes a drift banner down, and a consumer that treats it as "nothing to report" leaves a stale banner up after someone put the field back. A pod-template change is categorically different from the rest: in Kubernetes it CYCLES PODS, which makes it a deployment somebody made outside Ravion rather than a config edit. Those mint an `EXTERNAL` ModuleDeployment row; everything else stays here as report-only detail, because minting a deployment for an HPA scaling event would bury the deployment history under the cluster's ordinary breathing. properties: changeCount: description: >- How many changes the agent found, which can exceed `changes` when the list was truncated. format: int32 type: integer changes: items: $ref: '#/components/schemas/Deployments.EksWorkloadDriftChange' type: array changesTruncated: description: |- The agent bounds its change list; true means fields were dropped and the count below is the honest total. type: boolean drifted: description: >- False clears the finding. See the note above — this is a report, not a presence flag. type: boolean firstObservedAt: description: |- When this finding was FIRST observed, carried forward across every republish of the same finding. Without it a drift that has been standing for a week would read as new on every refresh. format: date-time type: string observedAt: description: When the stored finding was last refreshed by a report. format: date-time type: string referenceRevision: description: |- The Helm release revision whose stored manifest the comparison was made against (the agent reads it from the cluster's own release record), so a report against the spec a new deploy just installed is distinguishable from one against the revision before it. format: int64 type: integer rolloutCausing: description: |- True when at least one changed path is under the pod template, i.e. the change cycled pods. Recorded here as well as on the minted EXTERNAL deployment so the workload row can be read on its own. type: boolean sequence: description: |- The agent's monotonic ordering key for the DRIFT stream, independent of the rollout and regression streams. A report that arrived late after a reconnect must not resurrect drift the user already saw cleared. format: int64 type: integer required: - drifted - sequence type: object Deployments.EksWorkloadRegression: additionalProperties: false description: |- The regression half of an EKS workload's live state: the current answer to "was this workload healthy after its deploy and has it stopped being healthy". Absent on a workload that has never regressed. One object per workload, not one per episode. A flapping workload updates this object and raises `flaps` rather than accumulating episodes, which is what keeps the "one alert per episode, not one per flap cycle" guarantee alive past the control plane. properties: baselineAt: description: |- When the workload was last certified healthy — the "was previously healthy" half of the claim, stored so a user can check it. format: date-time type: string churnKinds: description: |- What the agent was tolerating when it decided to alert anyway — a node drain, a Karpenter consolidation, a scale event. Recorded because "we alerted despite a node drain" is materially different from "we alerted with nothing else going on", and because it is what makes a false positive diagnosable after the fact. type: string churnPodCount: format: int32 type: integer degradedAt: description: >- When the episode began. Two transmissions sharing it are the same alert, which is how a flap is told from a new outage — and it is what the delivered notification's `uniqueId` is built from, so an episode collapses onto one event rather than producing one per cycle. format: date-time type: string flapping: type: boolean flaps: description: |- How many times the workload recovered and degraded again inside this episode, and whether that count crossed the agent's threshold. format: int32 type: integer phase: description: >- opened | active | resolved. Free-form for the same version-skew reason every other agent-reported classification is: an agent newer than the control plane may report a phase this schema has never heard of, and a schema change is a far worse answer to that than a string. type: string resolution: description: |- recovered | superseded | workload_gone. Set only when phase is resolved — three different things to tell a user. type: string sequence: description: |- The agent's monotonic ordering key for the REGRESSION stream, which is independent of the rollout stream's `sequence`. A transmission that arrived late after a reconnect must be a silent no-op rather than a rollback — which for this object would mean re-raising an alert that had already resolved. format: int64 type: integer suggestedRemediation: description: >- restart | rollback. Free-form for the same version-skew reason every other classification in this file is. Unlike every other field on this model, this is NOT something the agent sent. It is the CONTROL PLANE's reading of the agent's report: a crash loop on a spec that has not changed since the baseline says restart, a regression that followed a deploy says rollback. Derived here rather than in the agent so the rule can be corrected by shipping the control plane, without a fleet-wide agent rollout. It is a SUGGESTION a human acts on. Nothing executes it — there is no automatic remediation path, and the reasons there is not are written down where the policy lives. type: string summary: description: Prose to lead an alert with. Nothing should parse it. type: string symptoms: description: >- The unexcused degradations, comma-separated in the agent's precedence order so the first entry is the cause rather than one of its consequences. type: string required: - phase - sequence type: object DeploymentResource.Ec2ScheduledEvent: additionalProperties: false description: Scheduled AWS maintenance affecting the instance. properties: code: type: string description: type: string eventId: type: string notAfter: $ref: '#/components/schemas/UtcDateTime' notBefore: $ref: '#/components/schemas/UtcDateTime' notBeforeDeadline: $ref: '#/components/schemas/UtcDateTime' type: object Deployments.AwsEc2InstanceReleaseGateStatus: description: State of the fail-closed Auto Scaling launch gate for one instance. enum: - PENDING - APPLYING - CONTINUE_REQUESTED - ABANDON_REQUESTED type: string Module.EcsCapacityProviderStrategyItem: additionalProperties: false description: >- Capacity-provider strategy entry. Mirrors AWS exactly. https://docs.aws.amazon.com/AmazonECS/latest/APIReference/API_CapacityProviderStrategyItem.html properties: base: description: |- Minimum number of tasks placed onto this provider before weighting kicks in. At most one entry in the strategy may set `base`. format: int32 maximum: 100000 minimum: 0 type: integer x-templateable: true capacity_provider: description: |- Capacity provider name (e.g. `FARGATE`, `FARGATE_SPOT`, or a custom EC2 capacity provider). type: string x-templateable: true weight: description: |- Relative share of tasks placed onto this provider when the strategy has multiple entries. format: int32 maximum: 1000 minimum: 0 type: integer x-templateable: true required: - capacity_provider type: object Module.EcsContainerOverride: additionalProperties: false description: >- Per-container portion of `EcsRunTaskOverrides`. Mirrors the AWS ECS RunTask `ContainerOverride` struct field-for-field — every override AWS supports for a container is expressible here. The `name` ties the override to a container in the task definition. https://docs.aws.amazon.com/AmazonECS/latest/APIReference/API_ContainerOverride.html properties: command: description: Replacement command — replaces the container's CMD array. items: type: string type: array x-templateable: true cpu: description: CPU units override for this container. format: int32 type: integer x-templateable: true environment: description: Environment variables merged on top of the task definition's. items: $ref: '#/components/schemas/Module.EcsEnvironmentVariable' type: array x-templateable: true environment_files: description: Environment files (S3 references) layered on top of `environment`. items: $ref: '#/components/schemas/Module.EcsEnvironmentFile' type: array x-templateable: true memory: description: Hard memory limit (MiB) override for this container. format: int32 type: integer x-templateable: true memory_reservation: description: Soft memory reservation (MiB) override for this container. format: int32 type: integer x-templateable: true name: description: Container name from the task definition. Required. minLength: 1 type: string x-templateable: true resource_requirements: description: GPU / InferenceAccelerator requirements override. items: $ref: '#/components/schemas/Module.EcsResourceRequirement' type: array x-templateable: true required: - name type: object Module.EcsEphemeralStorage: additionalProperties: false description: ECS ephemeral storage configuration properties: size_in_gib: description: Storage size in GiB (21–200 for Fargate, minimum 20) format: int32 maximum: 200 minimum: 20 type: integer x-templateable: true required: - size_in_gib type: object Module.EcsNetworkConfiguration: additionalProperties: false description: >- AWS ECS network configuration. Only the `awsvpc` shape exists on RunTask (host / bridge networking is task-definition-driven). https://docs.aws.amazon.com/AmazonECS/latest/APIReference/API_NetworkConfiguration.html properties: awsvpc_configuration: allOf: - $ref: '#/components/schemas/Module.EcsAwsVpcConfiguration' description: VPC configuration — required when set. x-templateable: true required: - awsvpc_configuration type: object Module.EcsPlacementConstraint: additionalProperties: false description: >- AWS ECS placement constraint. EC2/EXTERNAL launch types only — Fargate ignores these. https://docs.aws.amazon.com/AmazonECS/latest/APIReference/API_PlacementConstraint.html properties: expression: description: Cluster query expression — required when `type` is `memberOf`. type: string x-templateable: true type: description: |- Constraint type. `distinctInstance` spreads tasks one-per-host; `memberOf` filters by a cluster query expression. enum: - distinctInstance - memberOf type: string x-templateable: true required: - type type: object Module.EcsPlacementStrategy: additionalProperties: false description: >- AWS ECS placement strategy. EC2/EXTERNAL launch types only — Fargate ignores these. https://docs.aws.amazon.com/AmazonECS/latest/APIReference/API_PlacementStrategy.html properties: field: description: |- Attribute the strategy operates on (e.g. `instanceId`, `attribute:ecs.availability-zone`, `cpu`, `memory`). type: string x-templateable: true type: description: |- Strategy type. `random` ignores `field`; `spread` and `binpack` require `field`. enum: - random - spread - binpack type: string x-templateable: true required: - type type: object Module.EcsTag: additionalProperties: false description: AWS resource tag properties: key: description: Tag key type: string x-templateable: true value: description: Tag value type: string x-templateable: true required: - key - value type: object Module.EcsTaskVolumeConfiguration: additionalProperties: false description: >- Per-task volume binding for ECS-managed volumes (today: ephemeral EBS for Fargate). The `name` must match a volume entry in the task definition's `volumes` list. https://docs.aws.amazon.com/AmazonECS/latest/APIReference/API_TaskVolumeConfiguration.html properties: managed_ebs_volume: allOf: - $ref: '#/components/schemas/Module.EcsTaskManagedEBSVolumeConfiguration' description: Managed-EBS volume configuration. x-templateable: true name: description: Volume name — must match a `name` in the task definition. type: string x-templateable: true required: - name type: object Deployments.EcsEvidenceQuality: description: |- Evidence quality for the task's target-group snapshot. Computed once at task-stop by the selector. The snapshot is only persisted when the task's own target was observed, so the drawer never needs to distinguish "found" from "not found" — this label only separates the strong failure/healthy frame from a noisier initial/unused one. enum: - definitive - transient type: string Deployments.EcsTargetHealthCheckConfig: additionalProperties: false description: |- Health-check configuration AWS uses to probe the target group. Captured alongside `DescribeTargetHealth` so the drawer's target card can answer "what endpoint is AWS hitting?" — which is the first question readers ask when chasing an unhealthy target. properties: healthyThresholdCount: format: int32 type: integer intervalSeconds: format: int32 type: integer matcherSummary: description: |- Pre-formatted HTTP matcher summary (e.g. "200", "200-399"). Only meaningful for HTTP/HTTPS/GRPC. type: string path: description: Path AWS probes. Omitted for TCP-style health checks. type: string port: description: String because AWS accepts "traffic-port" as a literal. type: string protocol: description: |- "HTTP" | "HTTPS" | "TCP" | "TLS" | "GRPC" — the AWS enum verbatim. type: string timeoutSeconds: format: int32 type: integer unhealthyThresholdCount: format: int32 type: integer required: - protocol - port type: object Deployments.EcsTargetHealthSource: description: |- Which ingestion path wrote the snapshot. Used for debugging and UI labeling ("was this captured at task-stop or during the deploy poll loop?"). enum: - event-ingest type: string Deployments.EcsTaskTargetObservation: additionalProperties: false description: |- Projection of one `DescribeTargetHealth.TargetHealthDescriptions` entry. Carries the observable fields the drawer renders; omits volatile/internal bookkeeping. properties: description: description: AWS free-form failure description. type: string id: description: Display key "ip:port" or "container-instance-arn:port". type: string port: format: int32 type: integer rawId: description: |- Raw AWS target identity — private IP (awsvpc) or container-instance ARN (EC2). type: string reasonCode: description: AWS reason code (e.g., "Target.ResponseCodeMismatch"). type: string state: description: |- AWS target state: "healthy" | "unhealthy" | "draining" | "initial" | "unused" | "unavailable" | "unhealthy.draining". type: string required: - id - rawId - state type: object Deployments.EksWorkloadDiagnosisLog: additionalProperties: false description: >- What a container printed before it died, attached to the diagnosis that reports the death. `available` is the field this object exists for, and it must be read before anything else here. An unavailable capture and an empty one are OPPOSITE conclusions for whoever is debugging: "the process printed nothing" points at a crash before start-up, "we could not read the log" points at nothing at all and means try the next restart. So when `available` is false the text, the line count and `empty` are absent entirely rather than present-and-blank — there is nothing for a careless reader to render as silence. `exitCode` and `restartCount` are strings, not numbers, matching the agent's convention for every structured fact it reports: absent means "we do not know", which a number cannot express without a sentinel somebody eventually renders as -1. The capture's own exit code and restart count can DISAGREE with the diagnosis's. Under a capture rate limit the attached log may come from an earlier instance than the one the finding currently describes, and a reader has to be able to see that the logs are from restart 3 while the pod is on restart 40. properties: available: description: >- False means Operator could not read the log. See the note above — this is not the same as the container having printed nothing. type: boolean capturedAt: format: date-time type: string container: description: >- The container the output came from. Repeated here rather than inherited from the finding because a capture can hang off a pod-level finding, and a reader must never have to guess which container of a multi-container pod they are looking at. type: string empty: description: >- True when the container ran and printed nothing. Only meaningful when `available`. type: boolean exitCode: description: The run the output came from. Strings for the reason above. type: string explanation: type: string lines: format: int32 type: integer previous: description: >- True when the output came from the instance BEFORE the current one — the run that crashed, rather than the one running now. type: boolean reason: description: >- Why an unavailable capture is unavailable. Both empty when `available`. type: string restartCount: type: string terminationReason: type: string text: description: >- The captured output, verbatim. Present only when `available`. Bounded on the way in as well as by the agent: nothing a cluster prints may make one row's payload unbounded. type: string truncated: description: >- True when the output did not fit the byte cap. The END survives — the last thing a dying process said is the useful half — so the ellipsis belongs at the top. type: boolean required: - available type: object Deployments.EksWorkloadDriftChange: additionalProperties: false description: |- One field where the live object no longer matches what Ravion deployed. `previous` is ALWAYS the value Ravion deployed and `current` is ALWAYS what is running now — never the other way round, whichever direction the change went. Either may be truncated by the agent. properties: change: description: |- changed | added | removed. `added` means the cluster has a field the deployed spec does not, which is usually a mutating webhook rather than a person. type: string current: type: string path: description: |- Dotted path with list elements keyed by identity, e.g. `spec.template.spec.containers[api].image`. type: string previous: type: string required: - path - change type: object Module.EcsEnvironmentVariable: additionalProperties: false description: Environment variable as name-value pair properties: name: description: Variable name type: string x-templateable: true value: description: Variable value type: string x-templateable: true required: - name - value type: object Module.EcsEnvironmentFile: additionalProperties: false description: Environment file from S3 properties: type: description: File type enum: - s3 type: string x-templateable: true value: description: S3 ARN of the environment file type: string x-templateable: true required: - value - type type: object Module.EcsResourceRequirement: additionalProperties: false description: GPU or other resource requirement properties: type: description: Resource type — currently only GPU enum: - GPU - InferenceAccelerator type: string x-templateable: true value: description: Number of resources type: string x-templateable: true required: - type - value type: object Module.EcsAwsVpcConfiguration: additionalProperties: false description: >- AWS VPC configuration for `awsvpc`-mode tasks. https://docs.aws.amazon.com/AmazonECS/latest/APIReference/API_AwsVpcConfiguration.html properties: assign_public_ip: description: |- Whether the task ENI gets a public IP. Required for Fargate tasks pulling images from non-VPC-routable registries (e.g. public ECR) when there's no NAT. enum: - ENABLED - DISABLED type: string x-templateable: true security_groups: description: Security groups attached to the task ENI. items: type: string type: array x-templateable: true subnets: description: Subnet IDs the task ENI is placed into. items: type: string minItems: 1 type: array x-templateable: true required: - subnets type: object Module.EcsTaskManagedEBSVolumeConfiguration: additionalProperties: false description: >- AWS-managed Fargate ephemeral EBS volume configuration. https://docs.aws.amazon.com/AmazonECS/latest/APIReference/API_TaskManagedEBSVolumeConfiguration.html properties: encrypted: description: Whether the volume is encrypted at rest. type: boolean x-templateable: true filesystem_type: description: |- Filesystem type to format the volume with (`ext3`, `ext4`, `xfs`, `ntfs`). Defaults per OS family. type: string x-templateable: true iops: description: Provisioned IOPS — required for `io1`/`io2`, optional for `gp3`. format: int32 type: integer x-templateable: true kms_key_id: description: KMS key id (alias or ARN) for at-rest encryption. type: string x-templateable: true role_arn: description: |- IAM role ARN AWS uses to provision the EBS volume on the customer's behalf. Required. type: string x-templateable: true size_in_gib: description: Volume size (GiB). Range depends on `volume_type`. format: int32 type: integer x-templateable: true snapshot_id: description: Snapshot id to restore from. type: string x-templateable: true tag_specifications: description: Tag specifications applied to the EBS volume. items: $ref: '#/components/schemas/Module.EcsEBSTagSpecification' type: array x-templateable: true termination_policy: allOf: - $ref: >- #/components/schemas/Module.EcsTaskManagedEBSVolumeTerminationPolicy description: |- Termination policy — controls whether the volume is deleted when the task stops. x-templateable: true throughput: description: Provisioned throughput (MiB/s) — `gp3` only. format: int32 type: integer x-templateable: true volume_initialization_rate: description: |- Initialization rate (MiB/s) for snapshot-restored volumes — controls how fast blocks are pre-warmed. format: int32 type: integer x-templateable: true volume_type: description: |- EBS volume type (`gp3`, `gp2`, `io2`, `io1`, `sc1`, `st1`, `standard`). type: string x-templateable: true required: - role_arn type: object Module.EcsEBSTagSpecification: additionalProperties: false description: >- Tag specification for an ECS-managed EBS volume. https://docs.aws.amazon.com/AmazonECS/latest/APIReference/API_EBSTagSpecification.html properties: propagate_tags: description: |- Where the tags come from — explicit list, propagated from the task definition, or none. enum: - TASK_DEFINITION - SERVICE - NONE type: string x-templateable: true resource_type: description: |- Resource type the tags apply to. Currently only `volume` is supported. enum: - volume type: string x-templateable: true tags: description: Explicit tags. items: $ref: '#/components/schemas/Module.EcsTag' type: array x-templateable: true required: - resource_type type: object Module.EcsTaskManagedEBSVolumeTerminationPolicy: additionalProperties: false description: >- Termination policy for an ECS-managed EBS volume. https://docs.aws.amazon.com/AmazonECS/latest/APIReference/API_TaskManagedEBSVolumeTerminationPolicy.html properties: delete_on_termination: description: Whether to delete the volume when the task stops. type: boolean x-templateable: true required: - delete_on_termination type: object securitySchemes: BearerAuth: scheme: Bearer type: http ````