End-to-End Microflow Tracing on Mendix 9.24.25 with Dynatrace

0
Hi All,We are running Mendix 9.24.25 on-premises on a Windows Server with Dynatrace OneAgent already installed and configured for infrastructure-level monitoring. Currently we have the following working:- JVM CPU & memory metrics via OneAgent deep monitoring on the Jetty process- Database query performance (connection bus selects, deletes, commits)- HTTP request traces via IIS → Jetty service flow- Runtime health metrics (thread pool, GC, session counts)We are now looking to achieve end-to-end microflow-level distributed tracing — specifically, surfacing individual microflow names, execution durations, and call chains within Dynatrace traces.We explored the OpenTelemetry Marketplace module and were able to achieve this successfully. However, we have concerns about its long-term viability:- Only one version released (2023)- No updates since release- It is not an official Mendix moduleBefore committing to this module in production, we would like to understand if there are alternative approaches for Mendix 9.24.25 on Windows on-premises, such as:1. Any official Mendix-supported mechanism for OTel tracing on 9.x2. Any Dynatrace-native approach to capture microflow names without a Marketplace module3. Whether the built-in OTel runtime configuration (available in Mendix 10.x) has any backport for 9.24.x4. Any other community-recommended modules or patterns for microflow tracing in production on Windows on-premisesEnvironment Summary:Mendix Version : 9.24.25Hosting : Windows Server- OnPremisesAPM Tool : Dynatrace(OneAgent - Deep monitoring Enabled)Any guidance from the Mendix community would be greatly appreciated.
asked
2 answers
1

Hello,


Currently, Mendix 9.24.x does not have any official built-in OpenTelemetry support like Mendix 10.x. The OTel runtime configuration introduced in 10.x is not backported to 9.x.

From the Dynatrace side, OneAgent can monitor JVM, DB queries, HTTP calls, GC, etc., but it cannot automatically capture Mendix microflow names or call chains because those run inside the Mendix runtime layer.

For Mendix 9, the OpenTelemetry Marketplace module is honestly the most common approach people are using for microflow-level tracing. Even though it is not officially maintained by Mendix, many teams are still using it successfully for:

  • Microflow names
  • Execution duration
  • Distributed tracing

Other options are mostly custom solutions using Java actions + OpenTelemetry/Dynatrace SDKs, but that usually needs more maintenance effort.

As of now, there is no Dynatrace-native plugin or official Mendix-supported alternative for microflow tracing on 9.24.x.

If microflow observability is important long term, moving to Mendix 10.x would probably be the cleaner approach since OTel support is becoming platform-supported there.

Thanks !

answered
1

Hi Mahesh Seelam


On Mendix 9.24.25 there's no official path for microflow-level tracing — native OpenTelemetry runtime tracing only arrived in 10.18 (beta) and became fully supported in 10.24 LTS, with no backport to 9.x. Dynatrace OneAgent can't auto-discover microflow names on its own (they're a Mendix-internal abstraction, not JVM method boundaries), so you'd need an OTel agent regardless. That leaves the 2023 Marketplace module as effectively the only working option on 9.x — reasonable to pin and run short-term as low-risk tech debt since it's a JVM-level agent. But for supported, production-grade microflow tracing into Dynatrace, the real answer is upgrading to 10.24 LTS, where it's native and configured via JVM params exporting OTLP straight to Dynatrace.


But if you want I can connect and help you out for more info.


I hope this helps!

answered