Keep is an open-source AIOps and alert management platform written in Python that provides a centralized solution for managing alerts and incidents across multiple monitoring tools. The platform serves as a single pane of glass for alert aggregation, offering capabilities for deduplication, correlation, filtering, and enrichment of alerts from various sources. It enables bi-directional integrations with monitoring tools and supports customizable workflows for alert automation.
The repository demonstrates significant community engagement and active development. The repository shares contributors with other notable projects including twentyhq/twenty, mindsdb/mindsdb, and home-assistant/core, suggesting cross-pollination of ideas and practices within the broader open-source ecosystem.
Keep's feature set emphasizes alert intelligence and automation. The platform includes AI-powered capabilities for correlation and summarization, supporting multiple AI backends including OpenAI, Anthropic, DeepSeek, Ollama, LlamaCPP, Grok, and Gemini for enrichments and incident context gathering. The workflow automation functionality operates similarly to GitHub Actions but tailored for monitoring tools, enabling users to create sophisticated alert handling logic without leaving the platform.
The integration ecosystem is extensive, covering major observability and monitoring platforms. Supported tools include Datadog, Grafana, Prometheus-compatible systems, Elastic, Dynatrace, New Relic, Azure Monitoring, GCP Monitoring, CloudWatch, and numerous others. The platform also integrates with specialized monitoring solutions like AppDynamics, Coralogix, Checkmk, Centreon, and LibreNMS. This breadth of integration support positions Keep as a vendor-agnostic alert aggregation layer that can work across heterogeneous monitoring environments.
The repository is classified across multiple domains including Monitoring, Observability, Alerting, Incident Management, Health Checks, Metrics, Logs, Centralization, and Notifications, reflecting its comprehensive approach to alert management. The platform includes customizable dashboards and provides documentation at docs.keephq.dev, with a live demo available at platform.keephq.dev. The project actively welcomes community contributions and maintains a public issue template for provider integration requests, indicating a structured approach to expanding integration capabilities based on user demand.
