You can crawl Jive content outside using rest API for one time then
webhooks later or may be cron jobs. My suggestion would be to index space
information along with content. Then, you can use this space information
and access level of logged in user while making API call to Elastic search
to bring only those results. We recently built one such integration for our
On Wednesday, June 15, 2016, Andy Ebert <
1 person found this helpful
The key piece here and the most complex bit is determining permission levels to content in the search engine.
I know when we created a similar connector, creating a user access list for content items was a PITA. Users being added to or removed from security groups and social groups, changes to space hierarchies, content in user containers (with multiple authors)... ahhh! The mess that our code turned into trying to replicate the Jive permission model in our connector is still burned into my brain.
Just adding that this is the real issue. For reasons of performance as well it can become a real problem since this type of system must check permissions on each piece of content before rendering it to the user or risk breaking security.
I have to explain this all the time to audiences because everyone just assumes everything is searchable and they forget they don't have access to everything.
Thanks for sharing your experiences on implementing indexing that includes permissions. This aligns with the complexity that I had expected.
Looks like we might need to restrict the indexing to public content only - to fit this piece of work into the project scope - and think about indexing private content in a later phase.