What I like most about Grafana is its vendor-agnostic approach and overall flexibility. You’re not locked into a single ecosystem; instead, you can pull together metrics, logs, and traces from a wide range of data sources (Prometheus, CloudWatch, Elasticsearch, SQL databases) into unified, highly customizable dashboards. The visualization features feel mature, alerting is easy to set up, and the large open-source community provides pre-built dashboards for nearly every common workload or infrastructure stack. Review collected by and hosted on G2.com.

The biggest challenge for me is the learning curve. If you’re not already comfortable with query languages like PromQL or LogQL, building custom panels and drilling into specific metrics can feel pretty intimidating at first. Bringing on team members who aren’t deeply technical also takes some extra time, along with a bit of coaching to get everyone up to speed.

A smaller but still noticeable pain point is dashboard housekeeping. After the team has created dozens of dashboards and folders, keeping things organized—especially permissions—and avoiding duplicate or overlapping views can get messy. Setting up alert rules with multiple conditions can feel a little clunky too, which stands out because the rest of the interface is otherwise very clean. Review collected by and hosted on G2.com.