Lifecycle cleanup must avoid removing images required by active services. In this case, a developer bakes the production API key into the container image. “An image cleanup or retention process respecting deployment dependencies” would not satisfy that requirement; the supported action is “The secret remains exposed wherever that image is distributed”. Image layers are reusable artifacts and should not contain live credentials.

ACI exposes container logs/events to diagnose runtime failures. In this case, a developer bakes the production API key into the container image. “Container instance logs and events” would not satisfy that requirement; the supported action is “The secret remains exposed wherever that image is distributed”. Image layers are reusable artifacts and should not contain live credentials.

Infrastructure pull, network or configuration failure can occur before app logging. In this case, a developer bakes the production API key into the container image. “Inspect container group provisioning state and events” would not satisfy that requirement; the supported action is “The secret remains exposed wherever that image is distributed”. Image layers are reusable artifacts and should not contain live credentials.

External ingress publishes a managed endpoint to accepted public clients. In this case, an ACI container crashes repeatedly and the operator needs output from its main process. “Enable external ingress for the container app with the appropriate target port” would not satisfy that requirement; the supported action is “Container instance logs and events”. ACI exposes container logs/events to diagnose runtime failures.

Secrets should be injected securely rather than embedded in container images. In this case, an ACI container crashes repeatedly and the operator needs output from its main process. “Secure secret configuration or Key Vault references supported by the hosting service” would not satisfy that requirement; the supported action is “Container instance logs and events”. ACI exposes container logs/events to diagnose runtime failures.

Image layers are reusable artifacts and should not contain live credentials. In this case, an ACI container crashes repeatedly and the operator needs output from its main process. “The secret remains exposed wherever that image is distributed” would not satisfy that requirement; the supported action is “Container instance logs and events”. ACI exposes container logs/events to diagnose runtime failures.

External ingress publishes a managed endpoint to accepted public clients. In this case, a container group exists but reports provisioning failure with no application logs. “Enable external ingress for the container app with the appropriate target port” would not satisfy that requirement; the supported action is “Inspect container group provisioning state and events”. Infrastructure pull, network or configuration failure can occur before app logging.

Image layers are reusable artifacts and should not contain live credentials. In this case, a container group exists but reports provisioning failure with no application logs. “The secret remains exposed wherever that image is distributed” would not satisfy that requirement; the supported action is “Inspect container group provisioning state and events”. Infrastructure pull, network or configuration failure can occur before app logging.

Internal ingress restricts callers to the relevant environment/network scope. In this case, a container group exists but reports provisioning failure with no application logs. “Configure external ingress if public access is intended” would not satisfy that requirement; the supported action is “Inspect container group provisioning state and events”. Infrastructure pull, network or configuration failure can occur before app logging.

Event-driven scaling can use non-HTTP signals to adjust replicas. In this case, a Container Apps service needs incoming external HTTPS requests. “A Container Apps KEDA-based event scaling rule for the supported queue” would not satisfy that requirement; the supported action is “Enable external ingress for the container app with the appropriate target port”. External ingress publishes a managed endpoint to accepted public clients.

ACI exposes container logs/events to diagnose runtime failures. In this case, a Container Apps service needs incoming external HTTPS requests. “Container instance logs and events” would not satisfy that requirement; the supported action is “Enable external ingress for the container app with the appropriate target port”. External ingress publishes a managed endpoint to accepted public clients.

Infrastructure pull, network or configuration failure can occur before app logging. In this case, a Container Apps service needs incoming external HTTPS requests. “Inspect container group provisioning state and events” would not satisfy that requirement; the supported action is “Enable external ingress for the container app with the appropriate target port”. External ingress publishes a managed endpoint to accepted public clients.

Event-driven scaling can use non-HTTP signals to adjust replicas. In this case, a Container Apps workload has internal-only ingress and external customers cannot connect. “A Container Apps KEDA-based event scaling rule for the supported queue” would not satisfy that requirement; the supported action is “Configure external ingress if public access is intended”. Internal ingress restricts callers to the relevant environment/network scope.

ACI exposes container logs/events to diagnose runtime failures. In this case, a Container Apps workload has internal-only ingress and external customers cannot connect. “Container instance logs and events” would not satisfy that requirement; the supported action is “Configure external ingress if public access is intended”. Internal ingress restricts callers to the relevant environment/network scope.

Infrastructure pull, network or configuration failure can occur before app logging. In this case, a Container Apps workload has internal-only ingress and external customers cannot connect. “Inspect container group provisioning state and events” would not satisfy that requirement; the supported action is “Configure external ingress if public access is intended”. Internal ingress restricts callers to the relevant environment/network scope.

External ingress publishes a managed endpoint to accepted public clients. In this case, a background worker should scale on a queue depth rather than HTTP requests. “Enable external ingress for the container app with the appropriate target port” would not satisfy that requirement; the supported action is “A Container Apps KEDA-based event scaling rule for the supported queue”. Event-driven scaling can use non-HTTP signals to adjust replicas.

Internal ingress restricts callers to the relevant environment/network scope. In this case, a background worker should scale on a queue depth rather than HTTP requests. “Configure external ingress if public access is intended” would not satisfy that requirement; the supported action is “A Container Apps KEDA-based event scaling rule for the supported queue”. Event-driven scaling can use non-HTTP signals to adjust replicas.

Container Apps revision management supports changing active traffic versions. In this case, a background worker should scale on a queue depth rather than HTTP requests. “Activate/route traffic to the previous healthy revision as supported” would not satisfy that requirement; the supported action is “A Container Apps KEDA-based event scaling rule for the supported queue”. Event-driven scaling can use non-HTTP signals to adjust replicas.

Explicit revision traffic allocation enables staged rollout control. In this case, a queue-processing app stays at one replica despite thousands of messages. “Configure multiple-revision mode and supported traffic weights” would not satisfy that requirement; the supported action is “Queue scaler trigger, authentication and min/max replica settings”. Autoscaling depends on a valid event rule and credentials to read queue signals.

Minimum replicas set the floor below which scaling will not go. In this case, a queue-processing app stays at one replica despite thousands of messages. “A minimum replica setting of one keeps an instance active” would not satisfy that requirement; the supported action is “Queue scaler trigger, authentication and min/max replica settings”. Autoscaling depends on a valid event rule and credentials to read queue signals.

External ingress publishes a managed endpoint to accepted public clients. In this case, a queue-processing app stays at one replica despite thousands of messages. “Enable external ingress for the container app with the appropriate target port” would not satisfy that requirement; the supported action is “Queue scaler trigger, authentication and min/max replica settings”. Autoscaling depends on a valid event rule and credentials to read queue signals.

Minimum replicas set the floor below which scaling will not go. In this case, a Container App revision behaves badly and a prior revision is known-good. “A minimum replica setting of one keeps an instance active” would not satisfy that requirement; the supported action is “Activate/route traffic to the previous healthy revision as supported”. Container Apps revision management supports changing active traffic versions.

Event-driven scaling can use non-HTTP signals to adjust replicas. In this case, a Container App revision behaves badly and a prior revision is known-good. “A Container Apps KEDA-based event scaling rule for the supported queue” would not satisfy that requirement; the supported action is “Activate/route traffic to the previous healthy revision as supported”. Container Apps revision management supports changing active traffic versions.

Scale-to-zero depends on min replica count and applicable scaling trigger. In this case, a Container App revision behaves badly and a prior revision is known-good. “Set supported scale rules and minimum replicas to zero” would not satisfy that requirement; the supported action is “Activate/route traffic to the previous healthy revision as supported”. Container Apps revision management supports changing active traffic versions.

Autoscaling depends on a valid event rule and credentials to read queue signals. In this case, a new revision takes 100% traffic immediately despite a planned canary. “Queue scaler trigger, authentication and min/max replica settings” would not satisfy that requirement; the supported action is “Configure multiple-revision mode and supported traffic weights”. Explicit revision traffic allocation enables staged rollout control.

Event-driven scaling can use non-HTTP signals to adjust replicas. In this case, a new revision takes 100% traffic immediately despite a planned canary. “A Container Apps KEDA-based event scaling rule for the supported queue” would not satisfy that requirement; the supported action is “Configure multiple-revision mode and supported traffic weights”. Explicit revision traffic allocation enables staged rollout control.

Scale-to-zero depends on min replica count and applicable scaling trigger. In this case, a new revision takes 100% traffic immediately despite a planned canary. “Set supported scale rules and minimum replicas to zero” would not satisfy that requirement; the supported action is “Configure multiple-revision mode and supported traffic weights”. Explicit revision traffic allocation enables staged rollout control.

Container Apps revision management supports changing active traffic versions. In this case, a developer wants Container Apps replicas to drop to zero when idle. “Activate/route traffic to the previous healthy revision as supported” would not satisfy that requirement; the supported action is “Set supported scale rules and minimum replicas to zero”. Scale-to-zero depends on min replica count and applicable scaling trigger.

An Always restart policy is unsuitable for run-once tasks. In this case, a developer wants Container Apps replicas to drop to zero when idle. “The container group's restart policy” would not satisfy that requirement; the supported action is “Set supported scale rules and minimum replicas to zero”. Scale-to-zero depends on min replica count and applicable scaling trigger.

Explicit revision traffic allocation enables staged rollout control. In this case, a developer wants Container Apps replicas to drop to zero when idle. “Configure multiple-revision mode and supported traffic weights” would not satisfy that requirement; the supported action is “Set supported scale rules and minimum replicas to zero”. Scale-to-zero depends on min replica count and applicable scaling trigger.

Container Apps revision management supports changing active traffic versions. In this case, a Container App maintains one replica during idle periods. “Activate/route traffic to the previous healthy revision as supported” would not satisfy that requirement; the supported action is “A minimum replica setting of one keeps an instance active”. Minimum replicas set the floor below which scaling will not go.

Autoscaling depends on a valid event rule and credentials to read queue signals. In this case, a Container App maintains one replica during idle periods. “Queue scaler trigger, authentication and min/max replica settings” would not satisfy that requirement; the supported action is “A minimum replica setting of one keeps an instance active”. Minimum replicas set the floor below which scaling will not go.

An Always restart policy is unsuitable for run-once tasks. In this case, a Container App maintains one replica during idle periods. “The container group's restart policy” would not satisfy that requirement; the supported action is “A minimum replica setting of one keeps an instance active”. Minimum replicas set the floor below which scaling will not go.