The data required by AIOps is driven by your business goals. Values our customers receive from AIOps include:
- Correlation of information generated within a single domain or across the service delivery ecosystem
- Identification and prioritization of service-impacting issues
- Automation of data collection, mitigation, remediation and avoidance steps
VIA AIOps can ingest and enrich streaming data from devices, network equipment, and applications in real-time. AIOps ingests customer experience measures, metrics, events, logs, and traces needed to synthesize these insights from across service layers from any source. Let’s take a closer look at these data types and how they are used by AIOps to deliver this business value.
Customer experience measures can include:
- Technical measures of application experience such as late-arriving video frame count or MOS score;
- Records from customer interactions such as in-app problem reports, DownDetector.com metrics, or even customer care agent dialogs; and
- Proxy measures for customer experience such as RED metrics (Rate, Errors, and Duration), failed API calls, or request latency/service call response times.
Metrics: Provide insights into the technical operation and performance of an application and infrastructure and can deliver both leading and trailing indicators of issues that impact customer satisfaction when using the service
Events: Detect something at a specific point in time within the application or infrastructure. This data is enriched with additional information to understand the context of an event in complex, virtualized services.
Logs: Provide detail around system activity that is valuable when determining root cause and fix actions.
Traces: Identify how requests flow through a distributed application and service. This is valuable when investigating service experience issues and performance bottlenecks identified through metrics and event analysis.
Data is ingested from existing applications and tools via native mechanisms or directly from devices with VIA Connectors. VIA Connectors enable fast data onboarding, including data ingestion, parsing and enrichment, as well as definitions to turn the raw data into useful information. Connectors include: the VIA SNMP Trap Connector, VIA syslog Connector (ingestion of RFC5424 and RFC3164 syslog messages) and the VIA Kubernetes Connector (MELT data from Kubernetes environments). VIA Connectors allow your team to focus on understanding and acting on the insights in the data.
VIA Connectors enable data generated from disparate sources to be correlated and compared. Bringing data together is one of the key values delivered by AIOps. Other systems do not build relationships and deliver business value insights relating customer experience with service performance, event and fault data.
Individual entity events and metrics do not contain enough information to fully contextualize and prioritize issues. The same event has differing impact depending upon where it occurs in the overall service flow. Applying an overall customer experience measure enables identification of issues that have impact on the business. Using an overall service performance measure also helps avoid the problem of having ‘blind spots’ for subsystems that are not providing a data to AIOps, often because they are operated by a separate team, such as authentication servers and firewall devices. The customer experience measure can expose service delivery issues caused by any part of the service flow, even when that component is not providing data to the AIOps system.
To deliver maximum business value, AIOps triages issues to the correct fix agent and automates actions without human intervention. Actions include: data collection to enable fast human decisions or actions to mitigate an ongoing issue, to remediate an issue, or to avoid a negative impact before a customer even perceives it.
Automation actions depend on detailed information from log and trace data to identify where the action should be executed. Leveraging GenAI, VIA AIOps can identify similar past issues, learn from the human actions when troubleshooting, resolve those issues, and apply actions taken in the current Incident context. Large volumes of semi-structured and unstructured data from documentation to troubleshooting call transcriptions to SharePoint and Confluence repositories can be leveraged by AIOps to identify similar past Incidents and generate recommended actions for manual or automated execution.
VIA AIOps is designed to ingest this vast variety and quantity of data in real-time, enrich and correlate it, turning it into actionable insights. These insights drive actions that avoid, remediate or mitigate issues. This is why AIOps is necessary to deliver high customer satisfaction in a complex service delivery environment.
One response to “AIOps: Transforming Data into Operational Business Value”
An important message I would like to emphasize is that VIA AIOps goes beyond detection.
VIA AIOps automates responses to minimize downtime by proactively mitigating issues. For instance, it can reroute traffic during network congestion, trigger automated fixes such as restarting a failing cloud service, and learn from past incidents using GenAI to suggest solutions based on historical troubleshooting data.
Whether in finance, retail, healthcare, logistics, telecom, … VIA AIOps enables smarter decision-making, transforming raw data into tangible business value. For example, an online banking system utilizing VIA AIOps can detect unusual spikes in login failures, correlate them with firewall logs, and automatically suggest a security check before customers even report an issue. Similarly, in a manufacturing plant, VIA AIOps can monitor IoT sensors to predict equipment failures and trigger preventive maintenance, avoiding costly downtimes. In a healthcare system, it can forecast IT infrastructure failures that could impact patient data access and proactively allocate resources to prevent system outages. Likewise, in a logistics company, VIA AIOps can analyze past delivery delays and recommend alternative routes in real-time when detecting weather disruptions.
In the telecom industry, MOS (Mean Opinion Score), which rates call quality on a scale from 1 (poor) to 5 (excellent) is a key metric for assessing customer experience. VIA AIOps can correlate MOS with other critical data points, such as network metrics: increased packet loss and jitter on VoLTE (Voice over LTE) calls, affecting real-time voice transmission. Infrastructure data: high CPU and memory usage on core network servers handling call traffic, suggesting resource exhaustion. Application logs: repeated failed API requests to the call routing system, leading to call setup delays and failures.
And instead of waiting for customer complaints, VIA AIOps takes proactive actions to restore service quality. It can automatically reroute voice traffic through alternate, less congested network paths, allocate additional compute resources to overloaded call-processing servers, and trigger alerts to engineers, recommending permanent fixes such as scaling up network capacity in the affected region. By doing so, VIA AIOps ensures optimal service performance and a superior customer experience.
———
Nuestra audiencia hispano hablante también tiene derecho 🙂
Un mensaje importante que me gustaría enfatizar es que VIA AIOps va más allá de la simple detección.
VIA AIOps automatiza las respuestas para minimizar el tiempo de inactividad al mitigar proactivamente los problemas. Por ejemplo, puede redirigir el tráfico durante la congestión de la red, activar soluciones automatizadas como reiniciar un servicio en la nube que ha fallado y aprender de incidentes pasados utilizando GenAI para sugerir soluciones basadas en datos históricos de resolución de problemas.
Ya sea en finanzas, retail, salud, logística o telecomunicaciones, VIA AIOps permite una toma de decisiones más inteligente, transformando los datos en bruto en un valor empresarial tangible. Por ejemplo, un sistema bancario en línea que utiliza VIA AIOps puede detectar picos inusuales de fallos en los inicios de sesión, correlacionarlos con registros de firewall y sugerir automáticamente una verificación de seguridad antes de que los clientes siquiera reporten un problema. De manera similar, en una planta de fabricación, VIA AIOps puede monitorizar sensores IoT para predecir fallos en los equipos y activar un mantenimiento preventivo, evitando elevados tiempos de inactividad, que son muy costosos. En un sistema de salud, puede prever fallos en la infraestructura de TI que podrían afectar el acceso a datos de pacientes y asignar proactivamente recursos para evitar interrupciones. Asimismo, en una empresa de logística, VIA AIOps puede analizar retrasos de entregas anteriores y recomendar rutas alternativas en tiempo real al detectar condiciones climáticas no deseadas.
En la industria de las telecomunicaciones, el MOS (Mean Opinion Score), que califica la calidad de las llamadas en una escala de 1 (pobre) a 5 (excelente), es un indicador clave de la experiencia del cliente. VIA AIOps puede correlacionar MOS con otros puntos de datos críticos, como métricas de red: aumento en la pérdida de paquetes y jitter en llamadas VoLTE (Voice over LTE), lo que afecta la calidad de transmisión de voz en tiempo real. Correlacionarlo con datos de infraestructura: alto uso de CPU y memoria en servidores de red que gestionan el tráfico de llamadas, lo que sugiere un agotamiento de recursos. Y tambien correlacionarlo con
registros de aplicaciones: nos encontramos con fallos que se repiten en las solicitudes de API hacia el sistema de enrutamiento de llamadas, causando retrasos en el establecimiento de llamadas y fallos en la conexión.
Lo más interesante es que en lugar de esperar a recibir quejas de clientes, VIA AIOps toma medidas proactivas para restaurar la calidad del servicio. Puede redirigir automáticamente el tráfico de voz a rutas alternativas menos congestionadas, asignar recursos informáticos adicionales en los servidores sobrecargados que procesan las llamadas, y activar alertas para los ingenieros, recomendando soluciones, como la ampliación de la capacidad de la red en la región afectada.
De esta manera, VIA AIOps garantiza un rendimiento óptimo del servicio y una experiencia superior para el cliente.