Thanks for reporting as it asked! I think this is the first report we've gotten since we deployed that feature 🙂
In this case I can tell you it was a temporary connectivity issue on our server, and nothing to do with those particular search terms.
Trying the same recipe a second time simply gets "We haven't found anything."
Generally, if there is a failure, we still cache the (lack of) results to avoid putting pressure on the system in case there is an upstream outage or similar issue that retrying won't resolve. The cache is short at about ~5mins. We only show the "error screen" on the original request, subsequent requests will show the general "no results" one.
Ah, interesting -- oz logo works, but oz logos turns up empty. So the issue seems to be that the plural of the word isn't recognized as such.
Changing the query caused a cache miss, allowing it to retry. Nothing particularly wrong with this query.
Probably nothing specifically actionable for this ticket, but thank you for reporting your experience - I'll pass it off to our infra team to review.
