@arielassaraf Does your pattern and flow discovery features require predefinitions or changing my logs?

@shexman Hey :) No it doesn't, our algorithm handles log data as is. Simply send us your logs just like you generate them today, in 24 hours we will discover all your log patterns and 72 hours later we will finalize building your flows and start alerting you on flow anomalies. Hope that answers your question.

@coralogix always a pleasure to serve you @nassaraf :)

@navotvolk Thank you! Make sure you use our promo, it is valid also for existing Customers :)

@iamredlus @nivo0o0 @ Awesome job Lior! Just look at this beautiful anomaly: https://ph-files.imgix.net/78ec6... Is there anything prettier than that? well done! @arielassaraf @guykroupp

@sdizdar1 @coralogix @cloudhq Thank you! this is exactly the feedback that keeps us motivated :)

@gil_bahat That was one of our goals when we designed the system: saving you crucial time so you can focus on developing your product :)

@gil_bahat We work hard to provide the most cutting edge machine learning tailored to your systems. Many more of these are planned for the near future!

@philipkushmaro Well, when we looked at Log data we saw that it is actually not Big Data, it's just the same log patterns arriving millions of times only with different variables. So the first step of our algorithm is to cluster millions of log records back into their original patterns. Once we have those patterns we find connections between them to form sequences of patterns, which are the software’s normal flows. When we identify an abnormal behavior of a pattern or flow, we automatically notify our users allowing them to handle a few log records on a specific minute. That instead of drowning in millions of records and relying on searching and endless slicing and dicing.

@urieli17 Thank you! always a pleasure to find the needle in your haystack

@jodoron Hey! We currently support Java (via SDK, Log4J or Logback), Node.js (Via SDK, Winston or Bunyan), .NET (via SDK Log4NET or NLOG), Ruby/RoR and Python. Regarding 3rd party SW support we expose a REST API which is pretty simple to follow to send your data from anywhere via HTTPS and we also support rsyslog and SyslogNG to ship logs from any server. Hope that answers your questions. *Very soon we are addind FluentD and Logstash.

@otnilevi Awesome question. Most companies index and visualize your data and allow you to create user defined alerts and that’s great. But even small and medium companies today generate tens of millions of logs a day, and having a centralized place to search, slice and dice these logs is sometimes not enough. We automatically understand the log data structure in order to turn millions of log records back into their unique patterns, then find connections between those patterns to form the baseline flows of your software and alert you once they break. That way we solve the two main problems companies have with their logs today: 1) There are too many logs to handle 2) Companies are reactive instead of proactive.

Hey! @genadyo, The pricing goes by data packages and the desired retention period. Even though we provide all standard log management capabilities + our unique algorithms, we decided to start with a pricing that is similar to other companies in our market with 1 important improvement: While most companies define a daily limit and throttle your data once you exceed it (meaning log data loss) we NEVER throttle but instead we calculate weekly and monthly packages and notify you once you exceeded your monthly limit so you can decide whether you want to expand the packages or hold the data sending. More info on our pricing: https://coralogix.com/pricing/ And don't forget to use the promo 'PH2017' to get 30% off all packages (can be worth thousands of dollars a year)

Thank you @orgarfunkel !

@levy_josef thank you so much ! You know You are one of our most veteran customer ? from the Beta and now 2.0.

Hi @danbrahmy, thanks for your interest :) We implement many kinds of algorithms at Coralogix; these include Bloom Filters and HyperLogLog for membership heuristics, agglomerative tree analysis, multivariate distribution MLE/MAP, multiple time-series correlations, regressions, and more. Chaining these algorithms together enables us to reverse-engineer the types of log-records you have, construct the patterns and flows of your system and alert on anomalies. p.s. You may read about the HyperLogLog algorithm we're using and other technical details on our company blog: https://coralogix.com/process-bi...