We encountered some challenges while using Dynatrace. Although the initial setup was smooth, fine-tuning alert thresholds and custom metrics took some time. Another challenge was that Dynatrace charges based on host units, so we had to carefully plan our agent deployments. The licensing model is expensive. Additionally, the complexity of setup is an issue. While OneAgent and auto-discover services are powerful, the setup is more complex compared to other tools such as Prometheus and Grafana. These integrations are simple and basic, but Dynatrace setup requires more complexity based on the environment. For new users wanting to use Dynatrace, it is difficult. However, the AI-related solutions and metrics took us to the next level for identifying and fixing things. Dynatrace requires an agent for operation. OneAgent is powerful, but it is also resource-heavy. On lightweight nodes or older systems, the agent can slightly impact performance. If Dynatrace could implement a lightweight agent behavior, we could make things faster. Additionally, if Dynatrace could add a long-term retention policy so that we could store more data and find fine-grained details, that would help us. While Dynatrace managed edition supports on-premises deployment, the SaaS version depends on cloud connectivity. For highly regulated or air-gapped environments, setup and updates can be challenging. Although the initial setup is smooth, if someone wants to fine-tune it and fully understand the tool end-to-end, it could be tricky.
