One of the biggest obstacles Mendix developers face is users reporting performance issues.
A long running microflow, a slow loading page etc.
However the Mendix platform doesn't really offer easy ways to analyze degraded performance.
Therefore I would want being able to enable "trace" mode on a microflow, and if that is enabled.
Mendix automatically measures every microflow step and reports its duration back to me in a structured log message at the end of the microflow execution. Where the log message summarizes every step taken in the microflow :
###ms, [microflow_name][Action #1 caption],
###ms, [microflow_name][Action #2 caption],
###ms, [sub_microflow_name][Action #1 caption]
###ms, [microflow_name][Action #3 caption]
etc.
###ms, TOTAL DURATION
And if there is SQL being executed I would like it if there was some postgres features used to report on performance, I think there are things like: pg_stat_activity, pg_stat_user_tables, pg_stat_database and auto_explain options.
This feature should only be usable on localhost, and would require postgres as a database setting. And this last bit I'm not even sure if it is possible to use in that way.
I was not talking about logging but the new OpenTelemetry functionality: Monitoring Your Mendix Apps with an APM Tool | Mendix Documentation
You can filter which microflows you want to log and which not. You don't have to buy any tools for it, except that you of course need a (free) tool to analyze the data and a place to store the data.
I can turn on logging on ConnectionBus_Retrieve and a lot of other details via different default trace log nodes, however I cannot filter them to a specific microflow which causes enormous log flooding on large active apps and you can barely investigate the issue, hence the desire to be able to configure this at a microflow level. And perhaps the answer is indeed, you need to buy all these things outside the mendix platform. But that kinda beats the whole purpose of low code and an extremely opinionated stack.
Isn't that achievable with the new OpenTelemetry support? We use a monitoring tool with the new OpenTelemetry connection set up and for us, that helps a lot by looking into possible performance issues