Discover IBM Z OMEGAMON AI Insights 2.1.0 new features, improvements and use cases for Predictive AI and machine learning to detect anomalies and improve system performance, capacity and cost controls—without needing ML expertise.
GA Version 2.1.0 - November 11, 2024
(7min read)
In a Nutshell
IBM Z® OMEGAMON AI Insights is an AI Engine and a set of dashboards to enhance SREs experience (Site Reliability Engineers) with anomaly detection and alerting of key performance metrics for use cases of performance, capacity and cost control.
It leverages ODP (IBM Z® OMEGAMON Data Provider) to stream historical data, infuse it with AI Insights and is shipped as a new component of Tivoli Management Services 6.3.3.
OMEGAMON AI Insights 2.1 containers are available from ShopZ and FixCentral.
🆕Containerized Deployment & Security
Easy-to-deploy Docker containers for s390x Linux to the ELK stack, fully secured by default.
🆕Scalability & Performance
Now scale vertically to handle large data volumes. Disabled by default for simplicity.
🆕 CICS & Storage Scope Expansion
- IBM Z® OMEGAMON AI for CICS to detect hourly anomalies and significant divergence of CPU Time and Transactional Response Time at the region level for performance and protection of the transactional workload.
- IBM Z® OMEGAMON AI for Storage to display growth rates, linear projection of growth rates and times at the HLQ and Storage Group level to anticipate potential end of storage behavior.
✅ Smarter, Lighter AI Engine
50% lower CPU usage and 75% smaller database footprint thanks to rearchitecture.
Why Upgrade Now?
✅ Smarter Alerts
Start streaming individual contexts to receive alerts on key metrics and faster triage.
✅ Effortless Migration
As easy as it gets, start your migration now to a more performant, low overhead and secure environment with this new version.
⚠️ Binaries Deprecated
Previous binaries are no longer supported. Containers unlock full scalability and performance.
📦 How to Get the Update
What's Next
- Actively working: Extended scope for Db2, x86 support, Grafana UI and Alerting, Webhooks integration, Model Accuracy and 15m anomaly detection.
- Exploring: Extended scope for MQ, Customer Use Cases and Feedbacks, Situations and Events exploitation
Situations.
We want to hear from you!
Have you faced hidden performance issues? Curious how AI could help?
👉 Share your story on IBM Idea portal or request a demo today.
📖 Read how OMEGAMON AI helps you spot mainframe performance issues
More details...
⭐ New Features and Improvements
- [Breaking Change] AI Engine is now container-only for full scalability.
Migration is seamless — just stop previous processes and run docker-compose
- 🆕Scalable microservices and Docker containers for s390x Linux environment
- 🆕Seamless integration into environment to synchronize AI insights into Kafka or other customer database accepting TCP input
- Enhanced orchestration, GRPC communication for scalability and parallelization
- Improved security and TLS protocol deployment by default
- 🆕AI model and dashboard for CICS Use Case #1 – Analysis of Transactional Response Time at Region Level with hourly alerts
- AI model update and new dashboard for CICS Use Case #2 – Analysis of CPU Consumption at Region Level with hourly alerts
- 🆕New case study dashboard for Storage Use Case exploration with linear regression and growth rate projections
- 🆕Extended filtering capabilities down to the field level for a reduced footprint
- Enhanced performance of the AI Engine enabling vertical scalability
- Better log coverage and levels for readability and troubleshooting
🐞 Bug Fixes
- Email subject is now dynamic
- Forecast no longer impacted at the time of an ILM rollover
⚠️ Deprecation and Known Issues
- Slight increase in memory usage during initial training
- Occasional folder creation delay during first-time model training
- Forecasting may vary with highly irregular or sparse time series
- 1.1.0 Binaries are no longer supported — containers only for s390x Linux
Key Features from previous versions
- Historical data collection through OMEGAMON Data Provider
- Long term storage for analytics and machine learning into Elasticsearch
- z/OS, Network and JVM primary metric model and out of the box Kibana dashboard
- Pattern, seasonality and recurrent spike modelization
- Dynamic baselining, automated retraining and explainable output
- Severity scores and alerting rules setup for critical anomalies awareness
- Binaries packaging for s390x Linux (deprecated)
#monitoring, #ArtificialIntelligence(AI), #IBMZ, #OMEGAMON, #AnomalyDetection
@Jim Porell, @Anna Murray, @Fabien Gautreault