Move deactivation of observation due to missing forward model

[oyvindeide]

It seems a bit strange to deactivate an observation when fetching it if the forward model has failed, that means the observation knows too much. Moving that knowledge.

Issue Resolves #my_issue

Approach Short description of the approach

Pre review checklist

• [x] Added appropriate release note label
• [x] PR title captures the intent of the changes, and is fitting for release notes.
• [x] Commit history is consistent and clean, in line with the contribution guidelines.

Adding labels helps the maintainers when writing release notes. This is the list of release note labels.

[codecov-commenter]

Codecov Report

Merging #3341 (2424dda) into main (8633615) will increase coverage by 0.00%. The diff coverage is 16.66%.

@@           Coverage Diff           @@
##             main    #3341   +/-   ##
=======================================
  Coverage   64.96%   64.97%           
=======================================
  Files         616      616           
  Lines       48231    48219   -12     
  Branches     4331     4327    -4     
=======================================
- Hits        31335    31331    -4     
+ Misses      15454    15443   -11     
- Partials     1442     1445    +3     

Impacted Files	Coverage Δ
libres/lib/enkf/enkf_analysis.cpp	4.05% <0.00%> (-0.12%)	:arrow_down:
libres/lib/enkf/gen_obs.cpp	42.10% <0.00%> (+2.43%)	:arrow_up:
libres/lib/enkf/meas_data.cpp	67.48% <50.00%> (ø)
ert/ensemble_evaluator/builder/_prefect.py	89.11% <0.00%> (-1.04%)	:arrow_down:
libres/lib/enkf/block_fs_driver.cpp	71.92% <0.00%> (-0.59%)	:arrow_down:
libres/lib/res_util/block_fs.cpp	66.46% <0.00%> (-0.31%)	:arrow_down:

:mega: Codecov can now indicate which changes are the most critical in Pull Requests. Learn more

[oyvindeide]

test ert please

[oyvindeide]

Closing as this will be refactored in a different way