Description
Service outages continue to occur when IT is surprised by failures that could have been prevented. This webinar will demonstrate how ServiceNow Operational Intelligence enables you to see infrastructure anomalies in real-time before services fail, instead of waiting for monitoring tools to detect a failure after it happens.
Join us for this webinar to learn how to:
- Proactively analyze infrastructure performance and thresholds
- Detect anomalies that may indicate service outages
- Transform from a reactive to proactive approach
Please register to watch this on-demand webinar
Pricing and Packaging
Is all of this native capability in the ServiceNow ITSM offering, or is this an add-on product?
Operational Intelligence (OI) is a separately licensed ServiceNow ITOM application in the Jakarta release, and it requires the ServiceNow Event Management application. Predictive alerts feature is an enhancement to Event Management in the Jakarta release.
Is event correlation a part of Event Management, or does this need to be configured/programmed separately?
Event correlation is part of Event Management.
How is Operational Intelligence priced? How long is the retention of metrics?
OI is priced per Configuration Item (CI). We store 1 minute of aggregated data for 8 days, 10 minutes of aggregated data for 96 days, and 1 hour of aggregated data for 13 months.
Integrations
How are the metrics captured, and where does ServiceNow store these metrics?
Metrics can be collected from monitoring systems through connectors and sent to the Service instance. You can also use the REST API to push metrics. Metrics are stored in a database called MetricBase, which was developed as part of the Jakarta release.
How long does it take to integrate a monitoring tool?
If you use an out-of-the-box connector, it should not take more than 30 minutes to flow end-to-end operational metrics. For other integrations that don't have connectors, you can integrate using the REST API in about 2 hours.
Is there a SCOM plugin?
SCOM connector is available out of the box.
Are there connectors for HP OMi, UIM, Solarwinds, PRTG, Kaseya, LogicMonitor, O365, or Azure?
We are adding a lot more connectors in Kingston release, as well as afterwards. If there is no connector yet for one of your tools, you can use the REST API to push metrics.
Is OI positioning itself to compete with Splunk?
No, OI can ingest data from Splunk. ServiceNow ITOM applications are purpose-built to work with ServiceNow ITSM applications on the ServiceNow platform to eliminate service outages.
Setup
What needs to be configured for Operational Intelligence's anomaly detection and Event Management's predictive alerts?
For OI's anomaly detection to work, operational metrics need to be ingested using connectors or REST API. There is no need for configuration. The machine-learning starts detecting anomalies within 8 hours of receiving data, and the models improve over time. 2 weeks of data are needed to detect weekly seasonality.
Event Management's predictive alerts are generated using historical alerts stored in Event Management. Machine-learning starts working effectively with around 3-4 weeks of data.
Does OI only work for Infrastructure monitoring or also for application monitoring?
OI works for both infrastructure and application monitoring.
Do Operational Intelligence and Event Management's predictive alerts require a full CMDB to be implemented?
No, as long as you have CIs for which you want to add metrics.
Do you need discover the CIs, or can you start from 0 with ITOM?
OI can bind metrics to existing CIs or it can create CIs as part of metric ingestion
Don't you need Service Mapping configured as well for the anomaly and predictive?
You do not need Service Mapping to for Anomaly and Predictive Alerts.
How long do you need to collect/gather events for predictive alerts to be in place?
We need at least 30 days or more data to do predictive alerts. More data is always better for machine learning.
If we set the thresholds in here, do we still need to set the alerts up in the separate tools?
You do not need to setup alerts separately. You have ability to promote anomaly alerts to regular alerts
What are the typical dynamic thresholds?
CPU usage, memory usage, response time, etc.
For anomaly detection, does the dynamic thresholds feature use SPC (Statistical Process Control)?
- Yes.
Can you adjust the baseline in anomaly detection to lower the false positives? For example, I want an anomaly created when the metric is 5% over baseline and the metric is greater than 30% and has been going on for 5 minutes.
Today we do not have the capability to over-ride baselines, but this is planned for future release. The example would automatically be treated as an anomaly and an anomaly alert would be generated.
What is the maximum time resolution for event reporting and anomaly detection?
1 minute.
Reporting
How do I pull a high-level report showing all CIs in a given period? Will your product also provide CIO- and CTO-level overall dashboard showing the view as a more Business view instead of infrastructure status?
You can build such report using Performance Analytics.
