improve crash handling to prevent MDW from going down and blocking aescan

[nikita-fuchs]

Hey, a check of the deployment's logs could maybe prove me wrong on this, but it feels like certain erroring API calls cause the whole MDW to go down and restart, leading to aescan being fully disfunctional. Example: When I caused this error #1655 by selecting the corresponding keyblock in the keyblock selection, the whole aescan takes "time to recover" and is not usable during that period. screen cap:

deleteme.mp4

[sborrazas]

@nikita-fuchs middleware continues to respond fast, these 500s are not stopping the middleware from responding on the rest of the endpoints. Perhaps this is something that can be proposed to the aescan team

[nikita-fuchs]

alright, it seems to not reproduce anymore for now.

[nikita-fuchs]

Correcting myself, hanging still appears.