OpenSearch Community Meeting - 2023-1114

Chat log

15:01:53 From Alejandro Guida To Everyone:
Hi! Sorry I don’t have microphone
15:02:55 From Alejandro Guida To Everyone:
I need one like David’s one, haha
15:04:19 From Kris Freedain To Everyone:
Did you attend OpenSearchCon? Please take the survey
https://amazonmr.au1.qualtrics.com/jfe/form/SV_7PotmoUlux9zZBk
15:04:34 From Kris Freedain To Everyone:
OpenSearch 2.11.1 being tracked
https://github.com/opensearch-project/opensearch-build/issues/4161
Tentative release day of Nov 21
15:07:43 From Kris Freedain To Everyone:
Amsterdam user group launch event
Amsterdam OpenSearch Launch Event with Zeta Alpha (Nov 10, 2023)
15:08:21 From Kris Freedain To Everyone:
freedak@amazon.com
15:10:42 From Alejandro Guida To Everyone:
No, nobody pay my flight ticket from Argentina, jaja
15:13:35 From Nathan Boot To Everyone:
Reacted to “No, nobody pay my fl…” with :rofl:
15:14:00 From Kris Freedain To Everyone:
Reacted to “No, nobody pay my fl…” with :rofl:
15:16:34 From Kris Freedain To Everyone:
OpenSearch 2.11.0 Release Meetings · OpenSearch
15:18:25 From Alejandro Guida To Everyone:
I’m still playing with OpenSearch Data Prepper Anomaly Processor (I told Nathan about that last Community Meeting, and happily somebode answer my question in the Community Forum, thank Nathan for your help), and really I didn’t see another option like this one, because it is “simple stream processing Anomaly Detection”. The only problem for know (I already completed a Feature request) is that you loose the “learning” of the model when you restarts Data Prepper Docker. Some detections were really good. Maybe you don’t have many documentation about what is “happening” for the detections, but in general are really good. What I don’t know is really how much memory you need by pipeline, so sometimes Data Prepper Docker container stop working.
15:19:02 From Alejandro Guida To Everyone:
Sorry that I can only write, :frowning:
15:19:45 From Sam Herman To Everyone:
See you all, nice catching up. Will post some updates on Benchmark next time
15:20:28 From Alejandro Guida To Everyone:
Docker compose
15:21:15 From Alejandro Guida To Everyone:
But somebody from OpenSearch already told me that there is not persist between restarts, is not implemented. Every time that you restart Data Prepper, it start “learning” again from the metric logs that receive
15:22:10 From Alejandro Guida To Everyone:
But remember that I’m talking about Data Prepper Anomaly Processor, not OpenSearch mail Anomaly Detection Process
15:22:12 From Romain Tartière To Everyone:
Not a question, but a request: there are dozens of roles in OS and when mapping permissions to users it is very tricky (from my point of view) to decide which roles I need. A cheat-sheet of “roles for common use cases” would be really awesome.
15:22:26 From Kris Freedain To Everyone:
Was it this one? Data Prepper - Additional Information about Anomaly Detector Processor
15:22:42 From Alejandro Guida To Everyone:
Yes
15:22:56 From Alejandro Guida To Everyone:
Is really really interesting
15:23:08 From Alejandro Guida To Everyone:
I didn’t see that feature in other product
15:23:30 From Alejandro Guida To Everyone:
For me the idea of “anamoly stream processing” is really amazing, and is simple to configure
15:24:36 From Alejandro Guida To Everyone:
For now, I need to go, because here is half past 8 night, and my family is asking “when we are going to have dinner?” haha
15:24:40 From Alejandro Guida To Everyone:
Thanks a lot to all