Quick overview
This workflow runs daily to collect incident, deployment, metrics, log-volume, and GitHub code-churn history, builds per-service risk profiles, asks Anthropic Claude to predict which services are most likely to fail next, and creates preventive ClickUp tasks with WhatsApp alerts for high-risk predictions.
How it works
- Runs on a daily schedule (or manually) and loads configuration such as service names, API endpoints, and risk thresholds.
- Fetches historical incident logs, deployment history, APM/metrics trends, error-log volume trends, and GitHub commit activity via HTTP APIs.
- Builds a risk profile per service from these signals and detects recurring patterns such as weekday clustering and incidents occurring within 48 hours of deployments.
- Sends the per-service profiles to Anthropic Claude and parses a ranked JSON list of predicted failures with probability, risk level, failure mode, recommended actions, and a predicted window.
- Routes predictions by risk level to send an urgent WhatsApp alert and create a high-priority ClickUp task for high risk, create a standard ClickUp task for medium risk, or log low-risk predictions without action.
- Waits for the verification window, rechecks the incident log and ClickUp task status, and sends a WhatsApp confirmation if an incident occurred or a prevention-success summary if no incident occurred and the task was completed.
Setup
- Provide HTTP credentials for your incident log API, deployment history API, APM/metrics API, and log aggregator API used by the workflow’s HTTP requests.
- Add a GitHub API credential (or update the code-churn request and credential) and set the repo owner used in the commit search query.
- Add an Anthropic (Claude) API key using HTTP header authentication and confirm the model name used in the request.
- Add a ClickUp API credential and set the target ClickUp list ID for preventive task creation.
- Add WhatsApp Business Cloud credentials and set the WhatsApp phone number ID and recipient phone number for alerts and summaries.
- Update the service list, risk thresholds, time windows, and all API URLs in the configuration step before activating the workflow.