Could exceptions identify specific resource accessed?

I'm looking at app logs that perform many get_* calls, one of them failing, but I can't tell which, neither from kubeclient stack portion, nor the calling app:

[----] E, [2019-06-21T10:03:51.073003 #19344:590f4c] ERROR -- : [JSON::ParserError]: 765: unexpected token at 'This request caused apiserver to panic. Look in the logs for details.
'  Method:[block (2 levels) in <class:LogProxy>]
[----] E, [2019-06-21T10:03:51.073119 #19344:590f4c] ERROR -- : /usr/share/gems/gems/json-2.1.0/lib/json/common.rb:156:in `parse'
/usr/share/gems/gems/json-2.1.0/lib/json/common.rb:156:in `parse'
/opt/rh/cfme-gemset/gems/kubeclient-4.1.2/lib/kubeclient/common.rb:501:in `format_response'
/opt/rh/cfme-gemset/gems/kubeclient-4.1.2/lib/kubeclient/common.rb:315:in `get_entities'
/opt/rh/cfme-gemset/gems/kubeclient-4.1.2/lib/kubeclient/common.rb:206:in `block (2 levels) in define_entity_methods'
/opt/rh/cfme-gemset/bundler/gems/cfme-providers-kubernetes-339029025f99/app/models/manageiq/providers/kubernetes/container_manager/refresher_mixin.rb:67:in `block in fetch_entities'
/opt/rh/cfme-gemset/bundler/gems/cfme-providers-kubernetes-339029025f99/app/models/manageiq/providers/kubernetes/container_manager/refresher_mixin.rb:65:in `each'
/opt/rh/cfme-gemset/bundler/gems/cfme-providers-kubernetes-339029025f99/app/models/manageiq/providers/kubernetes/container_manager/refresher_mixin.rb:65:in `each_with_object'
/opt/rh/cfme-gemset/bundler/gems/cfme-providers-kubernetes-339029025f99/app/models/manageiq/providers/kubernetes/container_manager/refresher_mixin.rb:65:in `fetch_entities'
/opt/rh/cfme-gemset/bundler/gems/cfme-providers-openshift-eb55293f9470/app/models/manageiq/providers/openshift/container_manager/refresher.rb:30:in `block in collect_full_inventory'

This is an unfortunate interaction between the app (manageiq) doing client.send("get_#{entity[:name]}") in a loop:
https://github.com/ManageIQ/manageiq-providers-kubernetes/blob/bf6594110f8716f128a387e5a0817e89645d248e/app/models/manageiq/providers/kubernetes/container_manager/refresher_mixin.rb#L58-L67
and kubeclient defining method dynamically, so best you see in stack trace is generic get_entity:
https://github.com/abonas/kubeclient/blob/cb998b94a11ca271c1493d5d993804fd1c215a42/lib/kubeclient/common.rb#L207-L212

I know how to log better in the calling app, but I wonder if kubeclient could do something to make stack traces more informative?

I have some (ugly) ideas for dynamic methods like get_pod, but here's a harder challenge:
what about dynamic interfaces like .get(kind: ..., ..., metadata: {name: ..., }) that we want to add in #332 ? This is by definition a single method.

• One idea I have is not worrying about descriptive stack frames, but annotating the final exception object with details — what operation, what kind, what namespace/name (or just full URL accessed).

Start with lib/kubeclient/common.rb, especially format_response, get_entities, and define_entity_methods, and compare the caller loop in refresher_mixin.rb at lines 58-67. Determine how failures can retain operation, resource, or URL context, including the proposed dynamic interface in #332; done means the reported get_* failure exposes which resource access failed.