Toiling in insignificance: ‘83-‘20 | Toiling in relative obscurity: ‘20-‘24 | Shutting it down here ‘24-Now (Bsky: @thetizzo.com, Mstdn: @[email protected]
Pro tip: put a different fake birthday into every coffee shop payment system you encounter then sit back and collect your free or discounted drinks all year long!
The fact that @honeycombio handles high cardinality data out of the box is a complete game changer compared to every other observability tool that doesn’t.
@antculver Playing adult softball scratched this itch for me for a while. It’s definitely not baseball but still fun to get out and hit and run the bases.
As a Nuggets fan who lives in Denver, it’s always fun to find out how the Nuggets played on twitter from the national media members since I can’t watch them on TV myself @AltitudeTV@comcast
@xeraa Agreed. I’ve been definitely working on restructuring the documents our indexes and this has dramatically reducing the deleted documents in the issue. Reindex was definitely a stop gap that we were running once a quarter when load was low. Definitely not preferable
I've been working on scaling an Elasticsearch cluster for the last several years. I wrote down everything I've learned. Hope it's helpful for someone!
https://t.co/fvkgQw5Sjv
@xeraa I will definitely try _split out the next time I have to increase the shards in an index though. Would be cool if the docs I linked could reference _shrink and _split as alternatives. 😁
@xeraa Yeah I think was another consequence of our document structure being complex and writes taking a long time but we had issues w/our web processes being overloaded while waiting for writes to finish until we moved it to a queue and off the web request.
@xeraa In our use case we have some very write heavy indexes due to the way documents are constructed that we haven’t fixed so I don’t think the merging has a chance to help us. I will have to look into the only_expunge_deletes as I’m not aware of this setting.