MID Server and MID Server distributed cluster for Metric Intelligence
Summarize
Summary of MID Server and MID Server distributed cluster for Metric Intelligence
Metric Intelligence in ServiceNow requires at least one MID Server configured specifically for Metric Intelligence and organized within a MID Server distributed cluster. This cluster type exclusively supports Metric Intelligence operations and ensures efficient processing and anomaly detection of metric data.
Show less
A single MID Server supporting Metric Intelligence must be part of a distributed cluster, even if it is the only server. The MID Server processes metric data, detects anomalies, and transmits both processed batches and immediate anomaly information to the ServiceNow instance. It can also support other applications like Discovery and Agent Client Collector (ACC).
Key Features
- Automatic Activation: Enabling Metric Intelligence automatically activates the MID Server distributed cluster plugin.
- Cluster Configuration: If automatic cluster creation fails, manual configuration is available with Customer Support assistance.
- Application Support: The Metric Intelligence MID Server supports additional applications and is included by default in the ALL option for supported applications.
- Domain Support: When Domain Extensions Installer is active, metric processing occurs on a MID Server within the same domain as the business service; otherwise, the global domain MID Server is used.
Hardware and Software Requirements
- Hardware: For supporting up to 1.7 million metrics (e.g., 17,000 Configuration Items with 100 metrics each), a MID Server VM or host should have at least 24 GB memory and 8 CPU cores. Lower scale environments can use 12 GB memory and 4 cores.
- Software: Supported operating systems include Windows Server 2008 R2, 2012 R2, 2019 (both 32-bit and 64-bit), and Linux distributions like Red Hat Enterprise Linux 6.6+ and CentOS 6.6+, running the compatible MID Server version.
Configuration Requirements and Best Practices
- User Permissions: The MID Server service logon user must have read access to the SCOM database (OperationsManagerDW) for pulling data.
- Throttling Mechanism: Enabled by default to protect the instance from data bursts, ensuring no data loss while preventing overload. It can be disabled temporarily for troubleshooting via configuration parameters.
- Java and Memory Overrides: Configuration files (wrapper-override.conf and wrapper.conf) allow tuning of Java parameters such as garbage collection and IPv4 preference. Memory settings should be adjusted according to available RAM, with recommended initial memory of 1024 MB and maximum memory at 4096 MB or higher.
- IPv6 Note: Metric Intelligence MID Server does not support IPv6 and should be configured accordingly.
- Validation: Ensure the MID Server is validated and in a running state before use.
Practical Benefits for ServiceNow Customers
Implementing Metric Intelligence with a properly configured MID Server distributed cluster allows customers to efficiently process large volumes of metric data, detect anomalies in real time, and integrate metric insights with other ServiceNow applications. This setup enables scalable monitoring and proactive incident management, contributing to improved service reliability and performance.
Using Metric Intelligence requires at least one MID Server distributed cluster which contains a single MID Server that is configured for Metric Intelligence.
When using a single Metric Intelligence MID Server, that server must be added to a MID Server distributed cluster.
To support the specified throughput, create a distributed cluster with a single MID Server that meets the Metric Intelligence MID Server requirements. When using a virtual machine (VM), only a single MID Server can run that supports Metric Intelligence. When working with Agent Client Collector, a MID Server can also be connected to multiple agents (ACC-M and ACC-L).
A MID Server supporting Discovery can also be used with other applications.
For details on setting up Metric Intelligence, see Get started with Metric Intelligence.
MID Server distributed clusters
Activating Metric Intelligence (com.snc.sa.metric), automatically activates MID Server distributed cluster (com.snc.agent.distributed.cluster). A MID Server distributed cluster supports the Distributed MID Server cluster type which is used exclusively with Metric Intelligence.
If automatic creation of a MID Server distributed cluster fails, you can configure the cluster manually. For details, contact Customer Support.
Metric Intelligence MID Server
The Metric Intelligence MID Server supports the Metric Intelligence feature by processing data and detecting anomalies. The MID Server transmits batches of processed data to the instance. If an anomaly is detected, then information about the anomaly along with the raw data is sent immediately to the instance.
- Metric Intelligence as a supported application
- The Metrics capability
- Running the Metric Intelligence Metrics extension
The Metric Intelligence application that the Metric Intelligence MID Server is configured with enables you to add additional supported applications to the same MID Server. By default, Metric Intelligence is included in the definition of the MID Server ALL option. For information about modifying the behavior of the ALL option when selecting supported applications, see Configure applications included in ALL Applications.
If Domain Support - Domain Extensions Installer is activated, then you can configure a MID Server with the Metrics capability, per domain. In this case, metrics for a business service are processed on the MID Server that is in the same domain as the business service. Otherwise, a MID Server from the global domain is used.
- Is validated. For more information, see Validate a MID Server
- Is in running state
- Meets all software, hardware, and configuration requirements
| Requirement | |
|---|---|
One MID Server with:
|
Supports approximately up to a total of 1.2 million metrics (12,000 CIs with 100 metrics per CI). |
One MID Server with:
|
Supports approximately up to a total of 1.7 million metrics (17,000 CIs with 100 metrics per CI). |
| Software | Supported versions | Additional requirements |
|---|---|---|
| Windows | 32-bit and 64-bit versions:
|
32-bit or 64-bit version of the MID Server |
| Linux |
|
32-bit or 64-bit version of the MID Server |
| Configuration | Requirement |
|---|---|
| MID Server service logon user (required only for pulling) | Must be set to a user with read access to the SCOM database (OperationsManagerDW).
|
MID Server throttling mechanism
The MID Server throttling mechanism protects against data bursts to the instance. Data burst protection prevents overloading the instance, while ensuring that you do not lose any data.
The MID Server throttling mechanism is enabled by default. You may want to disable this mechanism when troubleshooting, to ensure that the mechanism is not causing system malfunction.
The throttling parameters are MID Server configuration parameters. Configure these parameters on the MID Server record, on the Configuration Parameters tab. When you add a configuration parameter, select the matching mid.oi parameter from the Parameter
name list.
- To disable this feature for metrics, set the
mid.oi.metric_token_bucket_enabledparameter tofalse. - To disable this feature for anomalies, set the
mid.oi.anomaly_token_bucket_enabledparameter tofalse.
Configuration Override
Use the ../agent/conf/wrapper-override.conf file to configure Java override parameters, such as the garbage collector and IPv4.
To configure the garbage collector, override the existing garbage collector parameter. For example, replace -XX:+UseParallelGC with -XX:+UseG1GC.
To force IPv4, add the following parameter, using the next available wrapper.java.additional number:
wrapper.java.additional.<number>=-Djava.net.preferIPv4Stack=true
wrapper.java.additional lines. Parameters 101 through 200 are reserved for ServiceNow out-of-the-box parameters. Only override those parameters by uncommenting the existing lines and
changing the values.Use the ../agent/conf/wrapper.conf file to configure memory for Metric Intelligence. Update the following parameters:
wrapper.java.initmemory=1024or higherwrapper.java.maxmemory=4096or higher
Optionally, if a server has additional RAM available, for example 16 GB RAM, you can replace 4096 with 8192 in wrapper.java.maxmemory=4096.