Problem Detection

Problem detection is a key capability of the Aisera AIOps platform. Based on AIsera LLM driven clustering and longitudinal behavioral analysis on temporal and ticket specific attributes, the system identifies application and infrastructure issue that need to be resolved and which might or may have already caused incidents and/or outages.

Problem detection works by:

1. Real time data ingestion: Aisera AIOps continuously monitors IT infrastructure and applications, collecting data, such as, alerts, logs, metrics, and real-time system performance indicators. This data goes into ticket clustering module which identifies the relevant clusters and anomalies in them. These anomalies are indicative of incidents and the source of the anomalies as determined by the Causal Graph point to the PRB that needs to be created.

2. Leveraging ITOps specific AiseraLLM: The platform employs IT Operations-specific Large Language Model to analyze the textual data within incident tickets, alerts and logs. Aisera LLM is adept at correlating between machine generated (alerts and logs) and human generated (user reported incidents) data and create corresponding clusters.

3. Ticket Clustering: The system then performs unsupervised clustering of the tickets (alerts, logs, incidents) based on output from the AiseraLLM analysis.

4. Longitudinal analysis: The platform employs long range time series analysis that incorporates seasonality to identify anomalies.