Administer > Monitor your network > Monitor cloud instances and VMs > Troubleshoot cloud monitoring

Troubleshoot cloud monitoring

This section details potential performance, usage, and polling issues you may encounter while monitoring cloud instances and VMs.

The Orion Platform tracks cloud events in the Alerts & Events > Events manager and displays alerts and events on the Cloud Summary page as well as Details pages for instances/VMs. For instances/VMs managed as nodes, check the Node Details page.

If an issue occurs on the cloud service side, explore the AWS Management Console or Azure Portal. Orion Platform issues may be related to cloud account configuration, throttling, or polling. Also, review cloud monitoring recommendations and requirements. Consult your system administrator as necessary.

If you contact SolarWinds Customer Support, make sure you have cloud account credentials available.

This section contains the following topics:

Common cloud monitoring issues

Issue: The Orion Platform cannot access a cloud service account.

If the Orion Platform cannot access a cloud account after entering credentials on the Add Cloud Account page, make sure you entered valid credentials. Consult your system administrator to confirm account configuration.

The Orion Platform supports VMs deployed via the Azure Resource Manager but not VMs created using a classic deployment model.

Issue: "You are not authorized to perform this operation."

This message appears if you lack adequate permissions to manage a cloud instance/VM. Check credentials and account configuration. Azure account requires the User Access Administrator role to manage account permissions and perform actions against VMs, such as stopping polling.

Issue: Data does not display in resources on Cloud Summary or Cloud Instance/VM Details pages.

Check credentials and account configuration. Navigate to the Cloud Summary and Cloud Instance/VM Details pages to determine if cloud services are throttling instances/VMs. For instances/VMs managed as nodes in the Orion Platform, check the Node Details page.

To ensure the Orion Platform can collect data using publicly available APIs accessed via HTTPS protocol, the Orion server must be configured to communicate with public services. Use the default setting — public — in community strings for polled devices.

Issue: Slowed performance when polling data.

Check the amount of cloud resources. For optimal performance, SolarWinds recommends the following limits:

  • Up to 10 cloud service accounts
  • Up to 1,000 instances/VMs to monitor
  • Up to 1,000 volumes to monitor
  • Up to 1,000 instances/VMs managed as nodes
  • Up to 1,000 Orion agents deployed on managed nodes

If you exceed recommended limits, polling may slow or be throttled by cloud services. Consider expanding hardware, server sizing, CPU resources, memory, etc.

Issue: "Error polling instance/VM. Error polling volume."

This message appears if one or more polling issues reduced the amount of events and alerts that display in the Orion Web Console. A description is included, along with the specific instance/VM or volume. The message appears if the following issues occur:

  • Error polling one or more metrics for an instance/VM
  • Access error when polling an instance/VM
  • Access error when polling a volume

Review the following potential issues and resolutions:

  • For metric polling errors, check the log file.
  • For cloud service access issues, check account credentials and permissions in the AWS Management Console, Azure Portal, and Orion Web Console. See Configure cloud accounts for the Orion Platform.
  • To ensure that the Orion Platform can collect data from cloud service APIs, make sure the Orion server uses the HTTPS protocol. Community strings for polled devices must use the default setting — public.

Issue: "Error polling instances for cloud account."

This message lists AWS instances that could not be polled or accessed.

Check account credentials and permissions in the AWS Management Console, as well as account information in the Orion Web Console. See Configure AWS for cloud monitoring.

Issue: "Error polling volumes for cloud account."

This message lists volumes that could not be polled or accessed.

Check account credentials and permissions in the AWS Management Console, Azure Portal, and Orion Web Console.

Issue: "Error polling VMs for cloud account."

This message lists Azure VMs that could not be polled or accessed.

Check account credentials and permissions in the Azure Portal, as well as account information in the Orion Web Console. See Configure Microsoft Azure for cloud monitoring.

Credentials required to add AD apps and/or deploy Orion agents to VMs may vary based on how Azure is configured for your organization. Ask your subscription administrator for details.

Issue: "Error polling auto-scaling groups for cloud account."

This message lists auto-scaling groups that allocate resources to match performance requirements could not be polled or accessed.

Troubleshooting: Check account credentials and permissions in the AWS Management console, Azure Portal, and Orion Web Console. See Configure cloud accounts for the Orion Platform..

Issue: Throttling applied at the account, instance/VM, or volume level. The instance/VM displays as Unknown - Throttling Error.

Cloud service APIs constantly check incoming monitoring requests against account limits; see Amazon API Gateway Limits and Azure Resource Manager. If polling requests exceed default or custom thresholds, cloud services return the HTTP status code 429 and apply throttling to block requests.

Cloud service throttling triggers default alerts in the Orion Platform that appear on the Cloud Summary and Cloud Instance/VM Details pages. Verify Orion Platform alerts and review cloud service usage plans. To prevent throttling issues, consider extending polling intervals and/or disabling API polling for specific resources. See AWS throttling and Azure throttling.

Issue: "AWS CloudWatch polling limit threshold exceeded."

AWS CloudWatch provides 1 million polling requests limit per calendar month at no charge. If the polling limit threshold is exceeded, verify Orion alerts, and review AWS usage.

To display estimated usage for the month, edit cloud account properties and review the CloudWatch API polling section. You can request a limit increase from AWS and/or disable CloudWatch polling at the account level. See AWS polling.

To avoid exceeding polling limits, consider toggling Auto Monitoring off when adding a cloud account to block polling for new instances launched and then discovered for an account. You can enable monitoring for individual instances later.

If you disable CloudWatch polling at the account level in the Orion Web Console, cloud services stop polling for metrics but continue to check resource status. After monthly free request limits are exceeded, costs are incurred for the remaining time frame.

Issue: "Azure polling limit threshold exceeded."

Azure provides 1 million standard API calls per calendar month at no cost and out-of-the-box alerts are included to alert you when polling limits are exceeded. Azure measures API requests according to the amount of data points returned; see Azure polling limits. Azure will charge a subscription or tenant account for overages, depending how Azure is configured.

To avoid exceeding polling limits, consider toggling Auto Monitoring off when adding a new cloud account to block polling for new VMs discovered for an account. You can enable monitoring for individual VMs later.

Issue: Azure volumes do not display many statistics.

Cloud services typically gather more data for instances/VMs than for volumes. The Orion Platform retrieves data from cloud services so data displayed for instances/VMs in the Orion Web Console may exceed data displayed for volumes.

If your organization uses both AWS instances and Azure VMs, you may notice that more information appears in the Cloud Volume Details resource for AWS instances. This is due to the fact that Azure does not provide individual metrics for Azure VM volumes. For example, Azure volumes always display the Up state because Azure does not yet provide volume status via API.

To retrieve more metrics, manage VMs as nodes in the Orion Platform.

Issue: Why do Azure volumes always display as in an "Up" state in the Volume Details resource?

Azure volumes always display the Up state because Azure does not yet provide volume status via API so the Orion Platform cannot retrieve and display that data.

Issue: The following SocketException message "An operation was attempted on something that is not a socket" displays in the logs.

This exception may occur if a large number of connections are open, which is common when more than 10 cloud accounts are polled simultaneously. Reduce the number of cloud accounts for optimal performance, as described in Cloud monitoring recommendations.

Issue: What does the SendAzureRequestsInBatch setting on the Advanced Configuration page control?

When the SendAzureRequestsInBatch option is enabled, the Orion Platform uses batch operations to poll Azure VMs. If a Microsoft Azure account includes over 1000 VMs, the Orion Platform may receive a Status Code 503 - Service Unavailable error, as recorded in the VMwareJobs log file stored in the SolarWinds\Logs\VIM\Jobs folder.

Per Cloud monitoring recommendations, limit the number of cloud VMs/instances to 1000 for optimal performance. Another option is to disable batch operations for Azure polling.

  1. In the Orion Web Console, navigate to the Global tab on the Advanced Configuration page (http://localhost/Orion/Admin/AdvancedConfiguration/Global.aspx).
  2. Disable the SendAzureRequestsInBatch setting.
  3. Increase the AzureOneTimeJobTimeout value from 00:03:00 to 00:06:00 to allow extra time to add Azure accounts when SendAzureRequestsInBatch is disabled.

Issue: The Azure Portal uses two terms for the same value: Tenant ID and Directory ID.

Tenant ID and Directory ID are the same value but are referred to both ways in the Azure UI and documentation. Azure also refers to the Client ID as the Application ID and Application Key. See Find cloud account credentials for details.

Issue: The number of Azure requests in the Edit Account wizard varies from the number of requests in the Edit Account wizard.

As described in Understand Azure polling limits, Azure uses datapoints to calculate the number of API requests; each request is equal to 1440 datapoints but individual API requests may consume more or less points. The Orion Platform displays the total of individual API requests in the Edit Account wizard so the number of Azure requests should be close but may not match exactly..

Issue: VMAN recommendations do not affect cloud instances/VMs.

VMAN recommendations that trigger actions in virtual environments such as vCenter do not impact cloud instances/VMs.

Azure messages

The following messages may appear when performing actions against Azure VMs.

  • "Authentication_Unauthorized."
  • "No subscription found in the context."
  • "Your Azure account does not have authorization to perform action 'Microsoft.Authorization/roleAssignments/write' over scope '/subscriptions/{guid}'."

These messages appear if an Azure account does not have permission to register the AD app that provides read access to the Orion Platform so it can poll for status and metrics. Typically, this occurs if an organization only allows administrators to register apps and an account is not linked to an administrator role. Ask the subscription administrator to either assign an administrator role to the Azure account or enable users to register apps. See Configure Microsoft Azure for cloud monitoring.

Credentials required to add AD apps and/or deploy Orion agents to VMs may vary according to how your organization uses Azure. Ask your subscription administrator for details.

Expand troubleshooting

To better manage and troubleshoot cloud instances/VMs and volumes:

Check the following logs that are related to cloud monitoring:

  • Business Layer logs:
    • C:\ProgramData\Solarwinds\Logs\CloudMonitoring
  • Polling jobs:
    • C:\ProgramData\Solarwinds\Logs\VIM\Jobs
    • Search for “Cloud” in the log files.
  • Collector logs:
    • C:\ProgramData\Solarwinds\Collector\Logs
    • Search for "Cloud" in the log files.