How did you set up monitoring for a new service? How do you store dashboards?

To set up monitoring for a new service, I usually start by defining key metrics that reflect the health and performance of the service: response time, error count, CPU and memory load, disk usage, as well as business metrics if available.

Next, I configure the collection of these metrics using tools like Prometheus, integrating exporters or built-in service metrics. For logs, I use centralized systems such as ELK or Loki.

After collecting metrics, I create dashboards in Grafana, grouping indicators by meaning: overall picture, errors, performance. It's important to make dashboards understandable and informative for different teams.

I store dashboards in a version control system (Git), using the export/import feature of Grafana JSON configurations. This allows tracking changes, conducting reviews, and quickly restoring dashboards if needed.