Investigate a failed deployment with the evidence in view.
Ask Mica about a failed deployment or unhealthy service. It can inspect your connected Microtica environments, deployment history and logs to help you decide what to do next.
- Demo project
- Mica
- Staging deploy failure
Sample Mica conversation. Asked why the catalog-service deploy to Staging failed, Mica checks the service, reads logs and compares versions. It finds that the new pods fail readiness because CATALOG_DB_URL is missing in Staging, then asks for approval to deploy the previous version back to Staging.
AI-assisted investigation
What you get
An investigation grounded in available environment, deployment and log data, with the next action kept separate from diagnosis.
What you bring
- A Microtica project with connected environments and services.
- Mica can only investigate the data available through those connections and your project access.
How it works
Know what happens before you start.
- 01
Choose the environment
Open an affected service, failed deployment or signal. Keep the investigation tied to the workload and environment you are trying to recover.
- 02
Ask a specific question
Ask why the deployment failed or what changed. Mica can inspect environment health, build history, deployment logs and application logs.
- 03
Check the proposed action
Review the evidence and verify the recommendation. If Mica prepares a supported deployment, build or scaling action, the interface asks you to approve or decline it.
Context beyond a chat message
Environment details, deployed versions and recent build or deployment logs help connect an explanation to the affected workload.
A visible change boundary
Supported deploy, build and service-resource actions use approval cards. You can inspect the target and decline an action instead of executing it.
“Microtica provided us with excellent guidance and support throughout the migration process, and the cloud platform has been a breeze to use since then. We are very thankful for the cost savings measures the platform provides.”
Access and control
Reading evidence and changing a workload are different actions
The assistant has diagnostic tools and supported write actions. The underlying cloud and Kubernetes integrations can have deployment permissions. Do not treat the complete platform connection as read-only.
Try it on one service.
Start free with a development project, or bring your setup to a demo.