convore.json/groups/devops-1/measure-everything/messages.json

[{"user_id": 275, "stars": [], "topic_id": 8061, "date_created": 1298387061.2489469, "message": "DevOps folks: what sorts of things are you measuring? What's the most interesting correlation  you've discovered while mining operational stats?", "group_id": 4827, "id": 179512}, {"user_id": 141, "stars": [], "topic_id": 8061, "date_created": 1298422026.331049, "message": "i'm more on the \u201cmeasure anything\u201d end so i'd like to hear some examples too.", "group_id": 4827, "id": 185308}, {"user_id": 141, "stars": [], "topic_id": 8061, "date_created": 1298422098.044441, "message": "for my tiny apps i'm only doing really high level user actions (high enough that i'm recording activity with goals in Google Analytics)", "group_id": 4827, "id": 185309}, {"user_id": 275, "stars": [], "topic_id": 8061, "date_created": 1298439005.935792, "message": "@kastner - can you comment on how your graphite is configured for apache? (mod_python, wsgi, etc) my graphite did not come with a sample .conf file", "group_id": 4827, "id": 185943}, {"user_id": 3868, "stars": [], "topic_id": 8061, "date_created": 1298500874.397547, "message": "We're using wsgi. It's a pretty standard config -- it uses some django bits :)", "group_id": 4827, "id": 192320}, {"user_id": 3868, "stars": [{"date_created": 1298512354.7237611, "user_id": 141}], "topic_id": 8061, "date_created": 1298502303.6761179, "message": "The most interesting correlation we've seen is the # of threads in our bug forum vs. code deploys. When we push something that is generating issues, it can spike *really* fast (sometimes faster than other, more conventional metrics)", "group_id": 4827, "id": 192739}, {"user_id": 275, "stars": [], "topic_id": 8061, "date_created": 1298598432.950886, "message": "@kastner: https://github.com/sivy/node-statsd", "group_id": 4827, "id": 204929}, {"user_id": 1170, "stars": [], "topic_id": 8061, "date_created": 1298612234.091738, "message": "@steveivy I figured there would be more clients from you!", "group_id": 4827, "id": 206145}, {"user_id": 277, "stars": [], "topic_id": 8061, "date_created": 1299275995.343888, "message": "@kastner is there a reason i'd get a counter value of .1 even if i set the sample rate to 1.0? it seems that you're doing some division, but i'm not sure why...", "group_id": 4827, "id": 273340}, {"user_id": 277, "stars": [], "topic_id": 8061, "date_created": 1299276015.59168, "message": "as an example: I send $ echo -n 'hello.world:1|c|@1' | nc -u localhost 8125", "group_id": 4827, "id": 273342}, {"user_id": 277, "stars": [], "topic_id": 8061, "date_created": 1299276445.9323089, "message": "so, that makes sense", "group_id": 4827, "id": 273386}, {"user_id": 277, "stars": [], "topic_id": 8061, "date_created": 1299276031.370126, "message": "and i get: stats.hello.world 0.1 1299275829", "group_id": 4827, "id": 273344}, {"user_id": 277, "stars": [], "topic_id": 8061, "date_created": 1299276438.8923709, "message": "hmm. I guess this is related to the 10 second buckets", "group_id": 4827, "id": 273384}, {"user_id": 277, "stars": [], "topic_id": 8061, "date_created": 1299281059.305248, "message": "yeah. issued a pull request", "group_id": 4827, "id": 273885}, {"user_id": 277, "stars": [], "topic_id": 8061, "date_created": 1299281086.746788, "message": "it's a bit awkward because java makes some things awkward", "group_id": 4827, "id": 273888}, {"user_id": 277, "stars": [], "topic_id": 8061, "date_created": 1299281091.0244009, "message": "but, works fine", "group_id": 4827, "id": 273891}, {"user_id": 275, "stars": [], "topic_id": 8061, "date_created": 1299281017.699275, "message": "btw - @apgwoz is working on a java client. :-)", "group_id": 4827, "id": 273882}, {"user_id": 16005, "stars": [], "topic_id": 8061, "date_created": 1303479799.3046429, "message": "@kastner - I think we are seeing similar results to what @apgwoz is seeing.  Basically, we want to use StatsD to sample a value at a slow rate, say at a one sample per minute rate.  When doing so, I'm getting results that seem to be off by a factor of 10.  I assume that this is due to the 10 second period munge/flush..  But what is the best way to compensate for that?  Change sample rate, or possibly multiply the value we're wanting to measure?  Or is tinkering with Graphite's data retention helpful here?", "group_id": 4827, "id": 781487}, {"user_id": 3868, "stars": [], "topic_id": 8061, "date_created": 1304472656.692409, "message": "There's a newish stat that should show up for all counters - it's a new directory stats_counts that is the raw value passed along. It doesn't play nicely with graphite's automatic rollup, and integral() calls across retention periods aren't useful at all. But it should be close to what you're looking for.", "group_id": 4827, "id": 905438}, {"user_id": 16005, "stars": [], "topic_id": 8061, "date_created": 1306418964.7845039, "message": "@kastner - stats_counts is giving us what we were looking for - thanks!  We had an earlier version of stats.js.", "group_id": 4827, "id": 1192060}, {"user_id": 3868, "stars": [], "topic_id": 8061, "date_created": 1306606792.665549, "message": "Awesome. There are a few patches on github for similar but slightly different things as well. I'm debating what to merge and what to leave out.", "group_id": 4827, "id": 1213555}]