Monitoring

The Monitoring page

It follows one cluster at a time, with measurements every 30 seconds. Choose the period for the charts: 1 hour, 6 hours, 24 hours, 7 days or 30 days. Refresh reloads the data right away.

At the top:

• Healthy services: how many services pass their health checks, checked every minute;
• Machines: how many are ready, how many unavailable;
• Worker machine CPU and Worker machine memory: what’s being used, and what your services have reserved.

Your services

The table gives, for each service: its health, its instances, its restarts over 24 hours and how long it has been running. A service that restarts often, or whose instances aren’t ready, is the first one to look at.

Usage by project

CPU and memory, project by project. Fransys (system) groups the tools that run the cluster.

Machines

Each machine: its role (control node or worker machine), its cores, its memory, how long it has been running, its 5-minute load average, and its state (Ready or Unavailable).

Public access and storage

The domains, the load balancer addresses and the storage reserved per project. Actual volume usage and traffic metrics (requests, response times) aren’t reported yet.

A service’s details

On a project’s diagram or list, a service’s status badge opens its details:

• Health checks: each check, its status and its message;
• Latest log lines of the service.

If the service’s last deployment failed, the details say so, with See the deployment. Full log and monitoring opens the service’s monitoring.

With your agent

A connected agent reads the same metrics, and a service’s logs and health checks. For example, ask it: “Read my API’s logs from the last hour and explain the errors.” See Connect an agent.

Coming soon

Alerts and incident history.