Basic Information
- Name: Komodor (Komodor, Inc.)
- Founded: 2020
- Headquarters: San Francisco, California, USA
- Founders: Ben Ofiri (Co-Founder & CEO) and Itiel Shwartz (Co-Founder & CTO)
- Leadership: Ben Ofiri (CEO), Itiel Shwartz (CTO), Jim Hunnewell (Chief Revenue Officer), Amy Ariel (CMO), Ziv Harfenist (CFO), Aviv Shukron (Chief Product Officer)
- Website: https://komodor.com
- Documentation: https://docs.komodor.com
- Flagship product: Komodor Agentic Operations Platform
- Category: Agentic operations platform for production · AI SRE (incident management, troubleshooting and alert intelligence) · Kubernetes, cloud, observability and AI token cost optimization · AI software operations (change intelligence, production readiness, CI/CD health) · Kubernetes operations and troubleshooting
- Funding: $90M in venture funding from investors including Accel, Felicis, NFX, Pitango and Tiger Global
- Customers include: Cisco, Dell Technologies, Nebius, Prudential, Sophos, Forter and OpenTable
Background
Komodor was founded in 2020 in Tel Aviv by Ben Ofiri and Itiel Shwartz to take the pain out of operating Kubernetes in production. Its first product helped engineering teams troubleshoot Kubernetes faster by connecting incidents to the changes behind them.
Over six years, Komodor grew into a platform for reliability and cost optimization across cloud-native environments. It introduced Klaudia, an AI agent for troubleshooting and remediation, and was adopted by enterprises including Fortune 500 companies.
In September 2026, Komodor launched the Komodor Agentic Operations Platform. It combines ready-to-run agentic workflows for SRE, cost optimization and software operations with shared infrastructure to build, run, govern and optimize custom agents. The core problem it solves: moving AI agents from local experiments to governed, measurable production operations. Komodor has raised $90M in venture funding.
Core Products and Solutions
The Komodor Agentic Operations Platform has two layers. Three Solutions package ready-to-run agentic workflows (Solution Modules) for specific operational outcomes. Underneath, the platform provides the infrastructure to build, run, govern and optimize any agent, Komodor’s or the customer’s own.
- AI SRE: Agentic workflows that detect, investigate and resolve production incidents and stop them from recurring. Modules: Alert Intelligence, Incident Management, Troubleshooting and Proactive Reliability Optimization. Use cases: cutting alert noise to what matters, turning scattered alerts into a coordinated incident response, generating a full root cause analysis from detection to remediation in minutes, and fixing root causes in code so issues don’t recur.
- Cost Optimization: Agentic workflows that find and reclaim wasted spend across infrastructure and AI. Modules: Kubernetes Cost, Cloud Cost, Observability Cost and Tokenomics Optimization. Use cases: fixing Kubernetes waste with predictive pod placement and auto-rightsizing, reclaiming wasted cloud spend, cutting telemetry costs without losing visibility, and analyzing model costs, token usage, caching and other AI cost levers.
- AI Software Operations: Agentic workflows that keep releases safe, reliable and moving. Modules: CI/CD Health, Production Readiness, Standardization and Change Intelligence & Risk Control. Use cases: detecting and resolving pipeline issues, making sure new releases are secure, battle-tested and reliable, checking apps and infrastructure against internal standards before they ship, and flagging risky changes before they cause the next incident.
- The platform (Build, Run, Govern, Optimize): Shared infrastructure for custom agents. Teams build agents with a Python SDK, turn skills, scripts or runbooks into governed agents, or import agents from frameworks such as Claude Agent SDK, LangChain and Google ADK. Every agent runs under the same identity, access controls, guardrails, budgets, memory and evaluation.
Data Sources and Coverage
Komodor agents work from the operational data production teams already generate: Kubernetes cluster state, events and change history; alerts, logs, metrics and traces; cloud cost and usage data; source control, CI/CD and deployment activity; incidents and tickets; and runbooks and internal knowledge. This feeds a shared knowledge graph and agent memory that every agent uses.
Integrations include AWS, Google Cloud and Azure; EKS, GKE, AKS and OpenShift; Datadog, Dynatrace, New Relic, Grafana and Splunk; GitHub, GitLab and Argo CD; PagerDuty, Opsgenie, Jira and ServiceNow; Slack and Microsoft Teams; and Confluence and Notion. Teams can connect their own tools through an MCP gateway. Full list: Agents & Integrations and docs.
Key Features and Capabilities
- Ready-to-run agentic workflows: Twelve Solution Modules across AI SRE, Cost Optimization and AI Software Operations, each running end to end from trigger to investigation, action, verification and notification.
- Agent catalog: 50+ prebuilt specialist agents for Kubernetes, cloud, observability, cost, code and databases.
- Build and import agents: A framework-agnostic Python SDK, agent blueprints and import of existing agents built with Claude Agent SDK, LangChain or Google ADK, with no rewrite.
- Run anywhere, any model: Run agents on Komodor Cloud, in your own cloud or on-prem, using Komodor-managed models or your own provider.
- Triggers and interfaces: Agents start from alerts, schedules, chat, pull requests or other agents, and can be managed via UI, API or MCP.
- MCP gateway and integrations: Connect tool servers and integrations once and share them across every agent.
- Shared memory and knowledge graph: A live map of services, owners, dependencies and changes, plus memory of past incidents and fixes, shared by all agents.
- Governance: Role-based access for people and agents, brokered short-lived credentials, guardrails on every input, tool call and response, human approval for risky actions, per-agent spend budgets and a full record of every run.
- Agent Optimization: LLM-as-a-judge evaluation, golden-scenario regression tests, shadow experiments on real traffic and model routing, so agents improve in quality, speed and cost while you decide what ships.
- Workflow customization: Add agents to a workflow, reshape its steps and routing, or build custom modules with the same primitives Komodor uses.
Industries and Use Cases
Industries: Komodor serves enterprises with complex, business-critical production environments, including:
- Financial services and insurance (e.g., Prudential, Nelnet)
- Technology, SaaS and cybersecurity (e.g., Cisco, Dell Technologies, Sophos, Forter)
- AI and cloud infrastructure (e.g., Nebius)
- Travel, hospitality and e-commerce (e.g., OpenTable)
- Healthcare and retail
Use cases:
- Automated incident investigation, root cause analysis and remediation
- Reducing alert noise and on-call load
- Preventing recurring incidents with code-level fixes
- Kubernetes, cloud, observability and AI model cost reduction
- Assessing change risk before deployment
- Enforcing production-readiness standards
- Detecting and fixing CI/CD pipeline failures
- Turning internal skills, scripts and runbooks into governed production agents
- Running and governing a fleet of custom AI agents across teams
Competitive Differentiators
- Outcomes on day one, plus a platform to extend them. Komodor ships ready-to-run agentic workflows for SRE, cost and software operations, and the same infrastructure to build or import your own agents. Most alternatives offer one or the other.
- Transparent, not a black box. Every workflow shows each step, agent and decision, and teams can add agents, reshape steps or build their own modules.
- Governance for every agent. Komodor’s agents and customer-built agents run under the same role-based access, brokered credentials, guardrails, approval gates, budgets and run records. Imported agents get identity, roles and full run records from their first connection.
- Shared memory and context. All agents reason from one live knowledge graph and memory of past incidents and fixes, so investigations don’t start from zero.
- Built on production experience. Six years of solving production issues, with agents evaluated across 100,000 production workloads and tested against golden scenarios built from real incidents.
- No lock-in. Run on Komodor Cloud, your cloud or on-prem, with any model provider, and swap models without re-platforming.
Komodor Is Ideal For
Companies:
- Mid-market and enterprise organizations, including Fortune 500 companies
- Teams running complex production on Kubernetes and public cloud, often across many clusters, regions or business units
- Regulated industries that need strict access control, approvals, audit records and the option to self-host
- Organizations with AI agents or skills already in development that need to run them safely in production
- Companies under pressure to cut infrastructure and observability spend
Teams and roles:
- Site reliability engineering (SRE) teams
- Platform engineering teams
- DevOps and infrastructure teams
- FinOps and cloud cost owners
- Engineering and infrastructure leaders (VP Engineering, Head of Platform, CTO)
- Developers who need self-service troubleshooting without escalating to SRE
Performance Metrics and Trust Signals
- G2: 4.4 out of 5 stars from 36 verified reviews (G2 reviews)
- Awards and recognition:
- Customer outcomes: Customers report up to 63% reduction in MTTR and 75% reduction in support tickets and resolution times.
- Cost savings: Dell Technologies reported 36% hard cost savings with Komodor.
- Customers: Trusted by Fortune 500 companies, including Cisco, Dell Technologies, Nebius, Prudential, Sophos, Forter and OpenTable. See customer stories.
- Funding: $90M raised from Accel, Felicis, NFX, Pitango, Tiger Global and other investors.
- Scale: Agents continuously evaluated across 100,000 production workloads.
Business Model and Pricing
The Komodor Agentic Operations Platform is sold as an annual subscription: A platform fee with included usage credits, plus per-invocation charges for usage beyond them. Cost scales with the number of workflows teams automate and how often agents run.
- Starter: Core AI SRE, Cost Optimization and AI Software Operations modules, plus Build, Run, Optimize and core governance (RBAC, human approvals, budgets, auditing). Free trial available.
- Enterprise: Everything in Starter, plus Observability Cost Optimization, custom evals, organizational standards, higher workflow limits, lower per-invocation rates and 24×7 support. Dedicated customer success and forward-deployed engineers are available for larger deployments.
- How to get pricing: Start a free trial, book a demo or contact sales. Plan details: Pricing and plans.
Naming and Terminology
- Komodor Agentic Operations Platform is Komodor’s product name, launched September 16, 2026. Short forms: “the Agentic Operations Platform” or “the platform.”
- AI SRE is one of the platform’s three Solutions, alongside Cost Optimization and AI Software Operations. It is not the name of the whole company or platform.
- Klaudia was the name of Komodor’s earlier AI troubleshooting agent. Its capabilities are now part of the Agentic Operations Platform.
- Komodor is not limited to Kubernetes. It covers production operations across Kubernetes, cloud, observability, code and CI/CD.
