This article is based on reporting originally published by devops.com in June 2019, covering Moogsoft’s version 7.2 platform release.
What Changed in Moogsoft AIOps 7.2
Moogsoft released version 7.2 of its AIOps platform with several enhancements aimed at reducing alert fatigue, improving workflow automation, and making AI-driven recommendations more transparent for IT operations teams. The update touches three distinct areas: workflow automation, alert clustering visibility, and individual-level performance analytics.
Workflow Engine: Automating Incident Response
The headline addition in version 7.2 is a dedicated Workflow Engine. According to Gabe Coelho-Kostolny, senior product management and product marketing leader at Moogsoft, the engine lets teams trigger automated actions directly from within the platform or in connected external systems. Supported actions include:
- Push notifications to communication and collaboration tools
- Automated ticket creation in ITSM platforms
- Custom remediation tasks that slot into existing incident management workflows
The goal is to close the gap between an AI-generated insight and the operational action that follows—without requiring a human to manually initiate each step.
Expanded Integration Ecosystem
Version 7.2 extends Moogsoft’s integration library with new connectors and proxy support for a broad range of monitoring and collaboration platforms. New additions include:
- New Relic Insights — pulling observability data directly into the AIOps correlation engine
- Microsoft Teams — enabling notifications and workflow triggers within team chat
- Proxy support for polling integrations covering: Zenoss, Zabbix, VMware vCenter, VMware vSphere, SolarWinds, Broadcom Spectrum, and SevOne
The proxy-based polling approach allows Moogsoft to collect data from these platforms in environments where direct agent installation may not be feasible—a practical concern in large enterprise estates with strict network segmentation.
Real-Time Alert Clustering Visualization
One of the persistent criticisms of AIOps platforms is the opacity of their machine learning models—teams are asked to act on recommendations they cannot inspect or tune. Moogsoft addresses this directly in 7.2 by adding tooling that lets IT operations staff:
- Visualize alert clustering algorithms as they run in real time
- Fine-tune clustering parameters when the default behavior does not match operational expectations
Coelho-Kostolny positioned this as a trust-building feature. When operators can see why alerts are being grouped together—and can intervene if the grouping looks wrong—they are more likely to act on the platform’s recommendations rather than second-guessing them.
Individual Statistics Dashboard
The release also ships an analytics dashboard called Individual Statistics. The dashboard is aimed at engineering managers and team leads who need to move beyond team-level aggregate metrics. It provides drill-down views that surface each team member’s workload and key performance indicators—useful for capacity planning, workload balancing, and identifying burnout risk before it becomes a retention problem.
Why DevOps Teams Are the Core AIOps Audience
Coelho-Kostolny noted a strong correlation between organizations that have adopted mature DevOps practices and those actively evaluating or deploying AIOps. The connection is structural rather than coincidental: modern DevOps pipelines—with continuous delivery, microservices, containers, and cloud-native infrastructure—generate an order of magnitude more telemetry events than traditional IT environments. The volume and velocity of those alerts quickly exceeds what human operators can triage manually, creating the conditions where AI-assisted correlation becomes operationally necessary rather than just aspirational.
The concern in these environments is not whether AI will replace IT administrators, but whether IT administrators can keep pace with the rate of change without AI assistance. The practical framing is augmentation: AIOps handles noise reduction and pattern recognition at scale so that human engineers can focus on decisions that actually require judgment.
Data Quality Remains the Hidden Prerequisite
Every AIOps platform depends on large volumes of accurate, well-structured data about the IT environment. Coelho-Kostolny acknowledged that data collection often turns out to be a harder problem than organizations anticipate when they begin an AIOps implementation. Incomplete or inconsistent monitoring coverage upstream will degrade the quality of AI-generated recommendations downstream—no platform can compensate for fundamentally poor input data.
Evaluating AIOps Claims: True AI vs. Rules Engines
With AIOps adoption approaching what Coelho-Kostolny describes as a tipping point in the enterprise, vendor marketing has intensified. He warned buyers to scrutinize what is actually under the hood. A meaningful distinction exists between:
- Genuine machine learning and AI — systems that learn from historical event data, adapt to environmental changes, and surface non-obvious patterns
- Rules-based engines dressed as AI — traditional ITSM correlation rules marketed under an AIOps label, which do not learn and do not generalize beyond what an administrator explicitly programs
The practical test: does the platform’s behavior improve over time with more data and without manual rule updates? If not, it is likely rules-based regardless of the marketing language used.
The Confidence Problem in AI-Driven Operations
Even well-implemented AIOps platforms introduce a risk that is qualitatively different from human error: when an AI system acts on a flawed assumption, it tends to do so consistently and at scale. A single human administrator making a wrong call affects one incident; an AI acting on a bad model can affect every incident simultaneously. This makes explainability and human override capability—exactly what Moogsoft’s clustering visualization feature addresses—essential requirements rather than optional conveniences.
IT teams evaluating AIOps platforms should ask vendors directly: what controls exist to inspect, override, and retrain the model when its recommendations diverge from operational reality?
Summary
| Feature | What It Does |
|---|---|
| Workflow Engine | Triggers automated notifications, ticket creation, and remediation tasks from within Moogsoft or connected external systems |
| New Integrations | Adds New Relic Insights, Microsoft Teams, and proxy polling support for Zenoss, Zabbix, vCenter, vSphere, SolarWinds, Spectrum, SevOne |
| Alert Clustering Visualization | Real-time view of clustering algorithms with fine-tuning controls to build operator trust |
| Individual Statistics Dashboard | Per-member workload and KPI drill-down for engineering managers |
For further reading on how AIOps intersects with observability and incident management, see the coverage of AppDynamics’ AIOps approach on devops.com.