Understand automated web service remediation workflows

When a web service fails in production, the response typically requires multiple teams and manual steps, and each handoff adds time to the outage. Automation orchestrator compresses this process into a single workflow that runs in minutes instead of hours.

In this use case, Event-Driven Ansible triggers an alert to automation orchestrator. Automation orchestrator routes the alert through a task agent for analysis, pauses for human approval, and then executes an Ansible Automation Platform job template to resolve the issue.

How the workflow responds to a service failure

The following example demonstrates an automation orchestrator workflow that remediates a failed NGINX web service detected by Prometheus Alertmanager.

Stage What happens Node type
Detect Prometheus detects that an NGINX service is returning HTTP 5xx errors. Alertmanager sends the alert to Event-Driven Ansible, which forwards it to automation orchestrator. Event-Driven Ansible trigger
Analyze A task agent receives the alert payload and queries recent application logs and container metrics through MCP server tools. It determines whether a service restart is safe or whether a deployment rollback is needed. task agent step
Approve The workflow pauses and notifies the on-call human team with the task agent analysis and recommendation. An engineer reviews the recommendation and approves or rejects the action. approval step
Remediate After approval, the workflow launches an Ansible Automation Platform job template that restarts the NGINX service on the affected hosts. Ansible Automation Platform job template step