Edit

Supported resources and Actions for Chaos Studio Workspaces

This article lists the Azure resource types that Chaos Studio Workspaces can target. For each resource type, it shows the Actions you can run and the Scenario templates that use them. Use it to check which faults are available for your resources before you build a custom Scenario. For example, to test a Service Bus outage, find Microsoft.ServiceBus/namespaces, pick an Action, and copy its actionId.

Important

Chaos Studio Workspaces and Scenarios are in public preview. Microsoft provides this preview "as is" and "as available," and it isn't covered by service-level agreements or limited warranty. Microsoft provides customer support for the preview on a best-effort basis. This preview isn't meant for production use. For more information, see the following articles:

Find Actions for a resource type

Find your resource type in the following table. The Scenario templates column lists the supported Scenario templates that use an Action on that resource type. You can use every Action in a custom Scenario.

Resource type Actions Scenario templates
Microsoft.ApiManagement/service Update backend None
Microsoft.Automation/automationAccounts Start runbook Zone Down (optional runbooks)
Microsoft.Automation/automationAccounts/runbooks Start runbook None
Microsoft.Cache/redis Reboot Zone Down
Microsoft.Cache/redisEnterprise Flush database, Force unlink Cache Stampede, Cache Stampede with Process Crash
Microsoft.Compute/virtualMachines CPU pressure, Hibernate, Physical memory pressure, Redeploy, Restart, Shutdown Compute Zone Down and its database variants, Zone Down, VM Hibernate, VM Maintenance Reboot, CPU Pressure, Physical Memory Pressure
Microsoft.Compute/virtualMachineScaleSets CPU pressure, Delete instances, Physical memory pressure, Shutdown, VMSS shutdown Compute Zone Down and its database variants, Zone Down, CPU Pressure, Physical Memory Pressure
Microsoft.DBforMySQL/flexibleServers Failover, Restart DB Failover Under Load, DB Restart Under Load, Cache Stampede, Cache Stampede with Process Crash
Microsoft.DBforPostgreSQL/flexibleServers Failover Compute Zone Down + PostgreSQL Failover
Microsoft.DocumentDb/databaseAccounts Offline region None
Microsoft.EventHub/namespaces Change event hub state Event-Driven Messaging Disruption, Dependency Blackout
Microsoft.LoadTestService/loadTests Start, Stop DB Failover Under Load, DB Restart Under Load
Microsoft.Network/loadBalancers Update load balancer admin state None
Microsoft.Network/networkSecurityGroups Apply NSG rule DNS Outage, Microsoft Entra ID Outage, Key Vault Public Endpoint Outage, Dependency Blackout
Microsoft.ServiceBus/namespaces Change queue state, Change subscription state, Change topic state Event-Driven Messaging Disruption, Dependency Blackout
Microsoft.ServiceFabric/managedClusters Availability zone down None
Microsoft.Sql/managedInstances Failover managed instance Compute Zone Down + SQL Managed Instance Failover
Microsoft.Sql/servers/databases Failover Compute Zone Down + SQL DB Failover
Microsoft.Sql/servers/failoverGroups Failover None
Microsoft.Storage/storageAccounts Planned failover None
Microsoft.Web/sites Kill process, Restart Cache Stampede, Cache Stampede with Process Crash

Faults inside an Azure Kubernetes Service (AKS) cluster, such as pod delete and network latency, aren't in this catalog. They run through the Chaos Studio extension for AKS. To shut down AKS nodes, use the scale set Actions. For more information, see AKS resilience testing with Chaos Studio Workspaces.

Use an Action in a custom Scenario

  1. Find the Action in the Action reference. Copy its actionId.
  2. Add the Action to a Scenario in the Scenario designer, or in a Microsoft.Chaos/workspaces/scenarios resource. For the resource structure and a Bicep example, see Create a custom Scenario.
  3. Set the parameters that the Action requires. Set a duration for Continuous and Cancelable Actions.
  4. Give the Workspace identity the recommended role, or a custom role with the same permissions. For more information, see Permissions and identity in Chaos Studio Workspaces and Create least-privilege roles.

The Workspace runs each Action against the discovered resources of its target type in the Workspace scope. To limit the resources, use resource filters or resource exclusions.

Action types

Each Action has one of three types. The type controls how the Action uses duration and what happens when the Action ends.

Type Behavior
Discrete Runs one operation, such as a restart or a failover, and then completes. It doesn't undo the operation.
Continuous Applies an effect for the Action's duration, and then stops it.
Cancelable Starts an effect and stops it when the duration ends or when you cancel the run. Some Actions have a parameter, such as RestoreOnStop, that controls what happens at stop.

Action reference

The tables in this section come from the Workspaces Action catalog. Each actionId has the form <action name>/<major>.<minor>. Parameters marked required must have a value. Other parameters are optional.

Application and integration

Action actionId Target resource types Type Parameters Recommended role
Update backend microsoft-apimanagement-updatebackend/1.0 Microsoft.ApiManagement/service Cancelable BackendId: name of the backend entity in the API Management service. Url: the invalid URL to set on the backend to simulate a failure. API Management Service Contributor
Start runbook microsoft-automation-startrunbook/1.0 Microsoft.Automation/automationAccounts, Microsoft.Automation/automationAccounts/runbooks Cancelable RunbookName: name of the runbook to run. Required when the target is an Automation account. RunbookParameters: runbook parameters as a serialized JSON dictionary. Automation Operator
Start (load test) microsoft-azureloadtest-start/1.0 Microsoft.LoadTestService/loadTests Discrete TestId: the load test to start. Load Test Contributor
Stop (load test) microsoft-azureloadtest-stop/1.0 Microsoft.LoadTestService/loadTests Discrete TestId: the load test to stop. Load Test Contributor
Kill process microsoft-appservice-killprocess/1.0 Microsoft.Web/sites Discrete ProcessName: process to kill. Leave it empty to kill the main worker process. InstanceIds: instances to target. Zones: availability zones to target, such as ["1"]. Website Contributor
Restart (App Service) microsoft-appservice-restart/1.0 Microsoft.Web/sites Discrete SoftRestart: true applies configuration settings and restarts only if necessary. The default, false, always restarts the app. Website Contributor

Caching

Action actionId Target resource types Type Parameters Recommended role
Reboot microsoft-azureclusteredcacheforredis-reboot/1.0 Microsoft.Cache/redis Discrete RebootType (required): AllNodes, PrimaryNode, or SecondaryNode. ShardId: shard to reboot, for clustered caches only. Redis Cache Contributor
Flush database microsoft-managedredis-flushdatabase/1.0 Microsoft.Cache/redisEnterprise Discrete DatabaseName: database in the cluster to flush. Azure Managed Redis Contributor
Force unlink microsoft-managedredis-forceunlink/1.0 Microsoft.Cache/redisEnterprise Discrete DatabaseName: database in the cluster to unlink. LinkedDatabaseResourceIds: resource IDs of the linked databases to remove from the geo-replication group. Azure Managed Redis Contributor

The Reboot Action works on clustered and non-clustered Azure Cache for Redis instances. For a non-clustered cache, such as a Basic or Standard cache, don't set ShardId.

Compute

Action actionId Target resource types Type Parameters Recommended role
CPU pressure microsoft-compute-cpupressure/1.0 Microsoft.Compute/virtualMachines, Microsoft.Compute/virtualMachineScaleSets Continuous PressureLevel: CPU pressure, 1–100 percent. Virtual Machine Contributor
Physical memory pressure microsoft-compute-physicalmemorypressure/1.0 Microsoft.Compute/virtualMachines, Microsoft.Compute/virtualMachineScaleSets Continuous PressureLevel: memory pressure, 1–99 percent. Virtual Machine Contributor
Shutdown microsoft-compute-shutdown/1.0 Microsoft.Compute/virtualMachines, Microsoft.Compute/virtualMachineScaleSets Cancelable GracefulShutdown: true shuts down gracefully. Zones: availability zones to shut down. Applies to scale sets only. Virtual Machine Contributor
VMSS shutdown microsoft-virtualmachinescaleset-shutdown/1.0 Microsoft.Compute/virtualMachineScaleSets Cancelable GracefulShutdown: true shuts down gracefully. Zones: availability zones to shut down. Virtual Machine Contributor
Delete instances microsoft-compute-deleteinstances/1.0 Microsoft.Compute/virtualMachineScaleSets Discrete Percentage: percentage of scale set instances to delete, 1–100. The service selects instances at random. Virtual Machine Contributor
Hibernate microsoft-compute-hibernate/1.0 Microsoft.Compute/virtualMachines Cancelable None Virtual Machine Contributor
Redeploy microsoft-compute-redeploy/1.0 Microsoft.Compute/virtualMachines Discrete None Virtual Machine Contributor
Restart (VM) microsoft-compute-restart/1.0 Microsoft.Compute/virtualMachines Discrete None Virtual Machine Contributor
Availability zone down microsoft-servicefabricmanagedcluster-azdown/1.0 Microsoft.ServiceFabric/managedClusters Cancelable Nodetype: node type to apply the fault to. Zone: the zone to bring down. You can bring down one zone at a time. Service Fabric Managed Cluster Contributor

CPU pressure and physical memory pressure are agent-based Actions. For the operating system and VM requirements, see Agent-based Scenario requirements.

Databases

Action actionId Target resource types Type Parameters Recommended role
Offline region microsoft-documentdb-offlineregion/1.0 Microsoft.DocumentDb/databaseAccounts Cancelable RegionName (required): display name of the region to take offline, such as East US. Cosmos DB Operator
Failover (MySQL) microsoft-mysql-failover/1.0 Microsoft.DBforMySQL/flexibleServers Cancelable RestoreOnStop (required): true tries to restore the original primary when the Action stops. Contributor
Restart (MySQL) microsoft-mysql-restart/1.0 Microsoft.DBforMySQL/flexibleServers Discrete None Contributor
Failover (PostgreSQL) microsoft-azurepostgresql-failover/1.0 Microsoft.DBforPostgreSQL/flexibleServers Cancelable FailoverMode (required): PlannedFailover or ForcedFailover. RestoreOnStop: true runs a planned restore to the original primary when the Action stops. SkipIneligibleServers: true skips read replicas and servers without high availability. Contributor
Failover (SQL) microsoft-sql-failover/1.0 Microsoft.Sql/servers/databases, Microsoft.Sql/servers/failoverGroups Cancelable ForceAllowDataLoss: true uses a forced failover, which can lose data. The default, false, uses a planned failover. SQL DB Contributor
Failover managed instance microsoft-sql-failovermanagedinstance/1.0 Microsoft.Sql/managedInstances Discrete ReplicaType: Primary (default) or ReadableSecondary. SQL Managed Instance Contributor

Messaging

Action actionId Target resource types Type Parameters Recommended role
Change event hub state microsoft-eventhub-changeeventhubstate/1.0 Microsoft.EventHub/namespaces Cancelable desiredState (required): Active, Disabled, or SendDisabled. eventhubs: event hubs to change. Azure Event Hubs Data Owner
Change queue state microsoft-servicebus-changequeuestate/1.0 Microsoft.ServiceBus/namespaces Cancelable desiredState (required): Active, Disabled, SendDisabled, or ReceiveDisabled. queues: queues to change. Azure Service Bus Data Owner
Change subscription state microsoft-servicebus-changesubscriptionstate/1.0 Microsoft.ServiceBus/namespaces Discrete topic (required): topic name. subscriptions (required): subscription names. desiredState (required): Active or Disabled. Azure Service Bus Data Owner
Change topic state microsoft-servicebus-changetopicstate/1.0 Microsoft.ServiceBus/namespaces Discrete topics (required): topic names. desiredState (required): Active or Disabled. Azure Service Bus Data Owner

Change subscription state and Change topic state are Discrete. They don't restore the original state. To restore it, add a later Action that sets desiredState to Active.

Networking

Action actionId Target resource types Type Parameters Recommended role
Apply NSG rule microsoft-networksecuritygroupv2-applynsgrule/2.0 Microsoft.Network/networkSecurityGroups Cancelable Rules: the NSG rules to apply. Each rule sets Access (Allow or Deny), Direction (Inbound or Outbound), Protocol, and address and port ranges. Network Contributor
Update load balancer admin state microsoft-network-updateloadbalanceradminstate/1.0 Microsoft.Network/loadBalancers Cancelable AdminState: Down, Up, or None. BackendPoolName: the backend pool. BackendAddressNames: addresses in the pool to change. Network Contributor

Storage

Action actionId Target resource types Type Parameters Recommended role
Planned failover microsoft-storageaccount-plannedfailover/1.0 Microsoft.Storage/storageAccounts Discrete None Storage Account Contributor

List Actions with the REST API

The Action catalog is also available through Azure Resource Manager. Use it to get the full parameter schema and the exact permissions that each Action needs.

  1. Run the following command. Replace <subscription-id> with your subscription ID. The example uses the West Central US region.

    az rest --method get \
      --url "https://management.azure.com/subscriptions/<subscription-id>/providers/Microsoft.Chaos/locations/westcentralus/actions?api-version=2026-08-01-preview" \
      --query "value[].{name:name, type:properties.actionType, targets:properties.supportedTargetTypes[].targetType}"
    
  2. Check the result. Each entry shows the Action name, its type, and its target resource types:

    {
      "name": "microsoft-documentdb-offlineregion",
      "targets": [
        "Microsoft.DocumentDb/databaseAccounts"
      ],
      "type": "Cancelable"
    }
    

Each Action also returns parametersSchema, supportedTargetTypes[].requiredPermissions, and recommendedRoles. The recommendedRoles property holds built-in role definition IDs. To build a least-privilege role from requiredPermissions, see Create least-privilege roles.

Report issues and request features

Azure Chaos Studio is developed in the open. To report a bug, request a feature, or ask a question about Workspaces, Scenarios, or the Azure CLI extension, open an issue in the Chaos Studio repository on GitHub. By filing an issue, you can track its progress and see requests from other customers.

Next steps