Monitor
Observability of cloud native usually includes logs, metrics, and traces. Logs and traces are not within the scope of this article. Let's see what abilities Horizon provides for Monitoring(metrics).
Horizon is a standardized cloud native delivery platform, so it is natural for us to choose the Prometheus monitoring stack as the foundation of our monitoring capabilities. Horizon itself does not deliver the Prometheus monitoring stack, so users need to deploy Prometheus in advance. We recommend using prometheus-operator to deliver the Prometheus monitoring stack in a declarative manner, which will integrate well with Horizon's template system.
Usage
We have built several monitoring dashboards with best practice based on Grafana for users to quickly experience and use. Next, let's see how to use it.
Enter the monitoring dashboard for a particular application instance.
- Select a application instance from the application instance list page.
- Click Monitoring from its menus.
The following dashboards will be shown:
- Pod - Resources monitor at the pod level.
- Monitor - Resources monitor at the container level.
- Serverless - If the load is delivered based on Knative, this dashboard can be used.
The following time periods are available for selection:
- 1h
- 3h
- 6h
- 12h
- 1day
- 3day
- custom - Customize the start and end of the time period.
'Auto refresh' is also supported with the following options:
- Close - Means closing the 'Auto refresh'.
- 30s
- 1 min
- 5 min
These standard monitoring dashboards can meet basic observability needs. If you have further monitoring needs, you can contact your administrator to easily add more monitoring dashboards.