Choose your metrics collection method

Here’s the summary, then pick what fits. For AWS, metric streams suit production and scrape jobs suit quick evaluation. For Azure, the serverless method works well in production, with Alloy for more flexibility. For GCP, use Grafana Alloy. If you run multi-cloud or want managed identity, Alloy handles everything. AWS scrape jobs and Azure serverless both have hands-on learning paths right here. For the rest, follow the docs. Once metrics are flowing, continue to the logs section. Metrics tell you something’s wrong. Logs tell you why.