Glossary

Ops Singularity glossary

Plain-English definitions of the platform terms used across this site, plus the AIOps industry terms behind them. If you have seen a word like Sentinel, ProcBot, Sherlock, OIAO, or MOP and wondered what it means, start here.

Ops Singularity is an autonomous AIOps platform. The terms below describe its agents, its operating method, and the industry concepts it builds on.

Platform & agents

Ops Singularity

An autonomous AIOps platform that unifies IT operations, incident management, Kubernetes monitoring, cloud cost control, and security threat detection, under one intelligence layer powered by Sentinel AI. It is operated by VWAVES Technologies Pvt. Ltd.

Sentinel (Sentinel AI)

Ops Singularity's central reasoning engine. Sentinel autonomously observes signals across every connected tool, correlates them, identifies the root cause of an incident, and orchestrates the resolution. It is the intelligence core that the rest of the platform, including ProcBot and Sherlock, runs on.

ProcBot

Ops Singularity's execution agent. ProcBot runs validated MOPs (playbooks) using Ansible and shell commands to remediate incidents, either fully autonomously when policy allows or after a human approves the action. The same ProcBot drives no-code workflow automation in DataByte, VisionWaves' data-engineering platform.

Sherlock

Ops Singularity's post-incident fix validation engine. After a fix is applied, Sherlock confirms the incident is genuinely resolved, scores how effective the procedure was, detects recurring issues, and feeds those learnings back to Sentinel so the system keeps improving. Sherlock also powers autonomous root-cause analysis in DataByte, VisionWaves' data-engineering platform.

The OIAO method

OIAO

Ops Singularity's closed-loop operations cycle: Observe, Investigate, Act, Optimize. Every incident flows through the same four phases, and the Optimize phase feeds its learnings back into Observe, so signal correlation gets sharper over time.

Observe, Investigate, Act, Optimize

The four phases of the OIAO loop. Observe ingests and normalizes signals (metrics, logs, traces, alerts, security events). Investigate performs automated root-cause analysis, maps blast radius, and selects the right MOP. Act executes the fix through ProcBot or routes it for human approval. Optimize validates the outcome through Sherlock and scores the procedure.

MOP (Method of Procedure)

A validated runbook that codifies exactly how to resolve a specific type of incident. Ops Singularity stores MOPs in a library, ProcBot executes them, and Sherlock scores their effectiveness so the best procedures rise to the top.

Operational pillars

ServiceOps

End-to-end application observability, tracing a request from the user all the way to the database across every service in between.

ClusterOps

Kubernetes and container operations: cluster health, pod lifecycle management, CrashLoop analysis, and automated remediation.

SecurityOps

Security operations and threat detection, with detections mapped to MITRE ATT&CK techniques and tactics.

DataOps

Data pipeline operations: job health monitoring, data-quality drift detection, and ingestion lag tracking. For full data engineering, ML, and pipeline operations, see DataByte, VisionWaves' integrated data platform.

FinOps

Cloud cost intelligence: real-time spend monitoring, cost anomaly detection, and rightsizing recommendations.

BusinessOps

Business-process operations, operations that understand the business they support, not just the underlying infrastructure.

Industry & technical terms

AIOps

Artificial Intelligence for IT Operations. Applying AI and machine learning to automate and improve IT operations such as monitoring, incident response, and root-cause analysis.

AMS (Application Management Services)

Ongoing managed operations for an enterprise's application estate. Federated AMS refers to orchestrating application management across multiple providers or business units from a single intelligence layer.

MTTR (Mean Time To Resolution)

The average time taken to resolve an incident, from detection to confirmed fix. A core measure of operations performance.

RCA (Root Cause Analysis)

Identifying the underlying cause of an incident rather than just its symptoms, so the same issue does not recur.

SIEM (Security Information and Event Management)

Tooling that aggregates and analyzes security events from across an environment to detect threats.

ITSM (IT Service Management)

The practice and tooling for managing IT services, including incident ticketing and change management.

Runbook

A documented, repeatable procedure for handling a specific operational task or incident. In Ops Singularity, runbooks are formalized as MOPs.

Blast radius

The scope of systems, services, or users affected by an incident. Mapping blast radius is part of the Investigate phase.

A2A orchestration (agent-to-agent)

Multiple autonomous agents, Sentinel, ProcBot, and Sherlock, coordinating across systems to resolve an incident end to end without a human stitching the steps together.

AIOps concepts

Alert noise

The high volume of low-value, duplicate, or non-actionable alerts operations teams receive. Industry estimates suggest only a small fraction of alerts require action; cutting alert noise is a core job of an AIOps platform.

Alert fatigue

The desensitisation that sets in when engineers are flooded with so many alerts that genuine incidents get missed or ignored. A direct consequence of alert noise.

Event correlation

Grouping related signals (alerts, events, metrics) from different tools into a single incident, so teams see one problem with full context instead of hundreds of disconnected fragments.

Deduplication

Collapsing repeated or identical alerts for the same underlying issue into one, so responders are not paged many times for a single problem.

Enrichment

Adding context to a raw alert, such as the owning service, recent changes, and topology, so a responder can act without hunting across tools for information.

Anomaly detection

Using statistical or machine-learning methods to flag behaviour that deviates from a normal baseline, often surfacing a problem before it becomes a hard failure.

Predictive alerting

Warning of a likely incident ahead of time, based on early signals and learned patterns, rather than only reacting after an outage has already happened.

Observability

The ability to understand a system's internal state from the data it emits, typically metrics, events, logs, and traces. Ops Singularity sits on top of your observability stack rather than replacing it.

Telemetry (MELT)

The metrics, events, logs, and traces (MELT) that systems emit and that monitoring and AIOps tools consume to understand health and diagnose incidents.

Incident

An unplanned disruption or degradation of a service that requires a response. The unit of work an AIOps platform is built to detect, investigate, and resolve.

Severity (SEV)

A rating of an incident's business impact, for example SEV1 for a critical outage down to SEV3 or SEV4 for minor issues. Severity drives how an incident is escalated and who responds.

Escalation policy

The rules that define who is notified, and in what order, if an incident is not acknowledged or resolved within a set time.

On-call

The rotation of engineers responsible for responding to incidents outside normal working hours. Reducing on-call burden is a primary goal of autonomous operations.

Postmortem (post-incident review)

A blameless review after an incident that captures what happened, the root cause, and the actions needed to prevent recurrence.

SLO (Service Level Objective)

A target for service reliability, for example 99.9% availability, that a team commits to and measures against.

SLA (Service Level Agreement)

A contractual commitment on service performance between a provider and a customer, often carrying penalties for breach.

Error budget

The allowable amount of unreliability, calculated as one minus the SLO, that a team can spend before reliability work takes priority over shipping new features.

Self-healing

Operations that detect and remediate issues automatically, without human intervention. Ops Singularity's closed loop, Sentinel investigating, ProcBot acting, and Sherlock validating, is a self-healing pattern.

AI concepts

Agentic AI

AI that can autonomously make decisions, plan a sequence of actions, and pursue a goal with minimal human intervention, adapting as conditions change. Sentinel, ProcBot, and Sherlock are agentic.

GenAI (Generative AI)

AI that creates new content, such as text, code, or summaries, from patterns in its training data. In operations it powers plain-English investigation and automatically written incident summaries.

LLM (Large Language Model)

An AI model trained on large amounts of text that can understand and generate human-like language. LLMs are what let engineers query operations in plain English.

Human-in-the-loop

Keeping a human in the decision path to approve or guide an autonomous action. Ops Singularity acts autonomously when policy allows and routes to a human when judgment is required.

Want to see these in action? Explore Sentinel AI or request a demo.