Open-source incident.io alternatives

Keep is an open-source platform for managing alerts and incidents. Teams that run several monitoring tools can bring their alerts into a single view, which saves them from jumping between dashboards, and then clean up the noise through deduplication, filtering, correlation and enrichment.

The platform connects to a long list of monitoring and observability products, and its integrations work in both directions so that data can be synced back to those tools. Workflows act like GitHub Actions for monitoring, letting you automate responses to alerts. AI features can correlate alerts and summarize incidents, using backends such as Anthropic, OpenAI, DeepSeek, Gemini, Grok or local models served by Ollama and LlamaCPP.

Keep is written in Python and the repository lists its license as Other. It positions itself as a customizable alternative to hand-built alert pipelines, with dashboards, documentation, a hosted demo and a request process for new providers. It suits SRE and DevOps teams dealing with alert fatigue across tools.

OneUptime is an open-source monitoring and observability platform that aims to replace a shelf of separate SaaS tools. It covers uptime monitoring, status pages, on-call and alerting, incident management, logs, traces, metrics, APM and error tracking in one application, and adds AI-assisted features such as helping find the root cause and proposing a fix.

Its README maps each module to the tools it can replace: Pingdom and UptimeRobot for uptime checks across websites, APIs, ports, SSL and DNS, StatusPage.io for branded public and private status pages, PagerDuty and Opsgenie for schedules, escalation policies and SMS, call, push or Slack alerts, Incident.io for declaring and triaging incidents with post-mortems, Datadog and New Relic for traces and dashboards, Loggly for logs, and Sentry for exceptions.

The Community Edition is open source under Apache 2.0 and free to self-host, and OneUptime Cloud offers a free-forever plan with no credit card. It is written in TypeScript. It suits small and mid-sized engineering teams that want one platform for reliability work rather than integrating many vendors.

HolmesGPT is open-source software, an AI agent, that investigates production incidents and works out their root causes. It is designed to work with any stack, including Kubernetes, virtual machines, cloud providers, databases and SaaS platforms, and it presents itself as an SRE agent that pulls signals together instead of leaving engineers to hop between dashboards.

It is a sandbox-level project of the Cloud Native Computing Foundation, first created by Robusta.Dev, with significant contributions from Microsoft. A newer operator mode runs in the background around the clock to spot problems before customers report them, addressing the gap that most AI agents wait for a person to spot trouble and start an investigation. Topics mention Prometheus, Slack and Jira integrations, ChatOps and incident management. It is written in Python.

HolmesGPT is licensed under Apache-2.0 and runs on your own infrastructure with the language model provider of your choice. It suits SRE and platform teams exploring AI-assisted incident response and alert triage.

Oncall is a calendar application built by LinkedIn specifically for scheduling and managing on-call rotations. It is aimed at operations and engineering teams who need to track who is responsible for responding to incidents at any given time, and it is designed to act as a source of dynamic ownership information for paging systems such as Iris.

As a calendar-focused tool, its core job is representing on-call schedules in a way that other systems can query programmatically, rather than functioning as a full incident-management platform on its own. It integrates into a broader on-call and alerting workflow where a paging tool consults Oncall's schedule data to determine who to notify.

Oncall is written in Python and requires a MySQL database, with setup instructions covering local development, Docker Compose deployment, and configuration through a YAML file; dummy data can be imported for testing. It is released under the BSD-2-Clause license, and the README notes some current limitations, including no built-in test-running mechanism documented and a need to rebuild to apply code changes in some setups.