4 ms·
Log Analytics at Uber
- warrenm 4y agoWonder how much it's costing Uber to build, maintain, and extend their custom solution over any of the extant options They've got a note on infra cost savings ...but that's a tiny portion of the overall costs of a system this large "There are about 10K Kibana dashboards saved by Uber engineers to analyze logs each month, in addition to many services calling ES endpoints directly to retrieve logs." There's a huge part of the problem. most likely: how many of those "10k Kibana dashboards" are duplicating effort? I've worked with 30+ TB/day ingest Splunk environments that suffer from none of the issues described in this post from Uber ... though they were all architected correctly: but you do have to actually architect your environment properly on the infra side to get performance on the user side Also - Uber's now introduced a non-replicatable skillset to their inhouse users: learn this ClickHouse thing after spending 8 years getting to know ELK Maybe it's "worth it" ... but I'd wager probably not -------- Disclosure - I work with Splunk full time, but am not here to praise its virtues over other offerings