Architecture
How does Run When work?
RunWhen's focal point is troubleshooting sessions, where users and Digital Assistants collaborate to run Tasks, analyze results, and plan next steps for effective troubleshooting.
From Scripts to Sessions
Community: For community Authors, the RunWhen platform acts like their CI/CD platform with tools to turn a broad range of contributions in to Digital Assistant-ready Tasks.
Setup: For platform teams, the RunWhen Local tool scans Cloud and Kubernetes accounts to match enterprise resources with Authors' contributions. It generates the required configuration automatically.
Troubleshooting: End users see the results when they start troubleshooting sessions where they collaborate with Digital Assistants to find root causes, remediate or escalate.
Authors and their scripts
RunWhen Authors in our community can mix and match a variety of programing languages using our Task file format. When Authors register a git repository, the Platform builds, tests and augments Tasks with metadata from our AI models. The metadata is ~10x the size of the original source code, effectively pre-computing thousands of scenarios where this code may be used.
Building the map
RunWhen Local (an in-cluster agent) identifies application frameworks, open source projects, platform components and infrastructure that match with Tasks in the library. Matches are sent to the RunWhen Platform to generate a Map of your environment that Digital Assistants follow.
Troubleshooting sessions
Troubleshooting sessions ("RunSessions") are the heart of the platform. In response to users' queries, alerts, webhooks, etc., Digital Assistants follow the Map to suggest Tasks and run them on RunWhen Local instances in your cluster. Tasks cover a broad range of diagnostics, triage and remediation capabilities and are chained together to get to a root cause, a remediation or escalation.
Ensuring Your Data Security Is Our Priority
Our unique architecture gives various deployment options to fit a wide range of corporate security standards. By separating open community content and sensitive enterprise data and ensures that no enterprise data gets uses in the generative AI process.
