Is it not searching/indexing pages or attachments? Ignoring the frustrating interface and query syntax it (in theory) should share some characteristics with elasticsearch since it uses lucene on the backend. In my experience it does an ok job of indexing pages and some file types. It would be nice if it just dumped the strings in binary files for indexing but I can see how that wouldn’t be very user friendly.
But it can only do ok when the indexing is working as it should. It can be easy to miss a stale index for a while because search continues to work. Then you’re waiting for it to catch up or even for a full reindex. The problems seem to get considerably more frequent with their HA products since you’re now distributing the search/index.
My impression from running a few different pieces of software and maintaining some smaller elasticsearch clusters is that search is pretty difficult to do well. Or even decent for that matter. Development, UX and operational support have to come together just right and when it’s just a feature of a product or a service instead of a core competency you get milquetoast and frustration.
But it can only do ok when the indexing is working as it should. It can be easy to miss a stale index for a while because search continues to work. Then you’re waiting for it to catch up or even for a full reindex. The problems seem to get considerably more frequent with their HA products since you’re now distributing the search/index.
My impression from running a few different pieces of software and maintaining some smaller elasticsearch clusters is that search is pretty difficult to do well. Or even decent for that matter. Development, UX and operational support have to come together just right and when it’s just a feature of a product or a service instead of a core competency you get milquetoast and frustration.