As opposed to traditional runbooks that are stale almost immediately, these scripts are provided by experts in an open source community who receive bounties and royalties when their scripts are run. There are defaults for troubleshooting in-house code (e.g. grep the logs for errors, file an issue), and you can import or build your own.

Instead of hunting through wikis, context-aware Digital Assistants respond to stackoverflow-style queries in the web UI, Slack, IDPs and other developer tools to suggest scripts to run. When they run scripts, they provide suggested next steps and detailed post-mortem reports.

Use community SLI/SLOs and built-in Error Budget Alerts, or your own observability stack alerts to call in Digital Assistants proactively. They use the output of each script as a search for the next one, iterating through the environment to collect diagnostics, find a root cause and propose a solution for engineers to review.

At the heart of the platform is the concept of “sessions” when scripts are being run. Engineers across different teams can collaborate in a session, and can invite different Digital Assistants to assist in suggesting and running scripts.

During a session, the participants (both user and Assistants) scripts are being recorded in a session report for escalation or post-mortem analysis. The report summarizes open and closed issues raised by each script as well as the full output for deep diagnostics of the environment.

Analytics for each session are tracked and can be exported to excel, gsheets, csv, etc to show time saved, people involved, resources involved and gaps in automation coverage where escalations were required.