Hello, @AravindhanV
I haven’t implemented this myself, but given that the OA installation isn’t possible, I’d probably look at this more as an external monitoring use case than a classic full APM setup.
A few ideas that seem worth exploring are Synthetic Monitoring for availability and key user journeys, polling Pega health/diagnostic endpoints for node and platform health, and custom metric ingestion into Dynatrace for things like requestors, queue processors, agents, or general system health. Log ingestion could also be useful if your Pega Cloud exposes relevant logs, alerts, or exports, and a custom extension might make sense if you want to package the polling and metrics collection in a clean way.
My guess is that the most realistic outcome here would be a solution with focus on availability, performance indicators, node health, background processing health, and queue statistics, rather than deep code-level observability.
The main limitations I’d expect are the lack of deep tracing visibility without OneAgent, the fact that JMX may not be practical in a managed cloud model, and that success would depend heavily on what your Cloud actually exposes through the APIs, endpoints, or log exports. Because of that, it may end up being more of a health + KPI + synthetic monitoring integration than a full observability.
So if anyone here has already done something similar, I’d be very interested to hear what worked best in practice 😀
Max Lopes
