What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 129 Stack Exchange communities.

For a discussion on a meta site where I'm a moderator, I'm in the process of compiling a few statistics.

You can only perform 6 searches within a 60 second window, please wait a moment and try again.

sigh

None of the usual workarounds such as using the API or the data dump or Google are usable because many of my searches involve deleted questions.

Please relax (not lift, but relax to a reasonable rate) the search rate limits, at least for moderators. We're trying to do our job here, and a handful of moderators per site aren't going to overload the servers.

I'd be satisfied with an alternative such as accessing deleted questions through the API or the data explorer.

share|improve this question
2  
Out of curiosity, what kinds of frequent searches are you performing, the inability to perform them of which is hindering your moderation abilities? –  Robert Harvey Jan 5 '12 at 22:24
1  
@RobertHarvey I'm a moderator on a beta site. There, handling flags (which indeed doesn't require many searches) is only a small part of being a moderator. –  Gilles Jan 5 '12 at 22:32
    
Curious. It's been days since I've had to do a search. Searching is not part of my normal workflow at all. –  Robert Harvey Jan 5 '12 at 22:33
3  
@RobertHarvey Well, here's the answer. This type of community building question — evaluating a question type — is fairly common on beta sites. Moderators are expected to participate, because they are supposed to be community leaders; and our ability to search deleted posts helps when we're trying to evaluate quality metrics. –  Gilles Jan 5 '12 at 22:50
    
Good idea. This has been biting me more and more often too. –  Shog9 Jan 6 '12 at 3:49

1 Answer 1

up vote 11 down vote accepted

Moderators will get double searches in the next build. The primary reason for the search throttle is I/O speed on our web tier, since we don't keep the entire Lucene index in RAM (risky, very risky - and possibly not at all stable). The amount of I/O it takes combined with previously low speed of the SATA drives in those servers meant I/O contention was a real factor.

We're taking a longer look at search now, possibly with an architecture change (a dedicated pair with the index in memory for example) that'll eliminate the need to throttle (well, much higher anyway - still blocking crazy scraping bots) with a slew of other improvements.

Stay tuned, we'll have some more around this soon enough.

share|improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .