Added JDBC tracing section in Java Observability

[davidhunglam]

Initial draft about enabling JDBC tracing in the Hana JDBC driver for CAP Java applications.

[beckermarc]

That is very complicated. I was expecting the configuration maybe also to be a little CAP-specific. These docs should actually be part of the HANA JDBC driver documentation from my perspective..

Could we try if the approach we discussed yesterday with data source properties works as well? That would be more CAP-specific (as you need to know where to place these properties in application.yaml) and also much simpler and platform independent.

[davidhunglam]

Hi @renejeglinsky, as parts of this documentation are rather CAP-unspecific (especially tracing activation via command line), I'm not sure if this the right part of our Capire to add this information to.

As Marc said, this should actually be in the HANA JDBC driver documentation, but that documentation is lacking severely in details, and users will surely contact us for guidance.

What do you think, should it better be placed elsewhere?

[renejeglinsky]

Hi @davidhunglam ,
if we can't link to docs that we deem appropriate, then it's fine to have the docs in capire. Especially in this case, where it's not too much.

Anyway we could provide feedback to the HANA colleagues about what we would expect. And if they can provide those docs, we can then remove it here and set just a link. Would you/Marc like to give feedback?

[rjayasinghe]

does this only work with absolute paths? Or would jdbctraces/trave_.log would write to to jdbctraces in the root of the application / container?

[davidhunglam]

Relative paths work as well. But the important thing is to use a directory the current user (in the cloud native buildpack that's cnb) has write access to. Otherwise nothing will be written silently, there'd be no error message.

[rjayasinghe]

Not sure if the JDBC trace configuration is something I wanted to persist in the application.yaml file. I would rather document the approach with environment variables and restaging the application, instead.

[davidhunglam]

I was not able to make it work with the environment variable only. Somehow the environment variable is only able to overwrite the existing entry from the application.yaml.

[davidhunglam]

Hi @davidhunglam , if we can't link to docs that we deem appropriate, then it's fine to have the docs in capire. Especially in this case, where it's not too much.

Anyway we could provide feedback to the HANA colleagues about what we would expect. And if they can provide those docs, we can then remove it here and set just a link. Would you/Marc like to give feedback?

Hi @renejeglinsky, we've talked internally about it, and we've decided that our Capire might be the better place for the documentation. The Hana driver documentation is kept strictly environment agnostic, so they don't really care about where the driver runs.

[renejeglinsky]

Hi @davidhunglam , if we can't link to docs that we deem appropriate, then it's fine to have the docs in capire. Especially in this case, where it's not too much.
Anyway we could provide feedback to the HANA colleagues about what we would expect. And if they can provide those docs, we can then remove it here and set just a link. Would you/Marc like to give feedback?

Hi @renejeglinsky, we've talked internally about it, and we've decided that our Capire might be the better place for the documentation. The Hana driver documentation is kept strictly environment agnostic, so they don't really care about where the driver runs.

Thanks! :)

[rjayasinghe]

I think we can take this a first iteration.

[BraunMatthias]

Are we fine with the current proposal?

[renejeglinsky]

@smahati Could you review this before the next release?