try to catch more exceptions during OpenSearch indexing

[naknomum]

Expected Behavior
A user does something which triggers OpenSearch indexing (creating or modifying an Encounter) but something goes awry with OpenSearch -- the user's action should still complete, if possible.

Current Behavior
User does the thing, but OpenSearch exception bubbles up to browser and action is only partially completed.

Testing Notes
It is extremely difficult to replicate the behavior that causes OpenSearch "to go awry"; so this will be the challenge. The incident which triggered the ticket was a mismatch between OpenSearch docker image version and pom.xml configuration (which manifested as a json parsing problem). Good luck future-jon.

[TanyaStere42]

make this into code guidelines because this is now an NFR WildMeOrg/wildbook-docs#98

[naknomum]

created a PR in docs to address best practices here, so closing this ticket.

OpenSearch has now been through several iterations, so general cases of exceptions are now mostly being caught (and often dumps stacktrace). this narrows down improvements and future development to cases covered in the docs: basically try to prevent invalid data from derailing the entire document's indexing -- stop it with code before it can be an exception.