kaisecheng

Results 26 issues of kaisecheng

Logstash keystore init container slows down the startup process. To update the keystore, `logstash-keystore` command starts and stops jvm for each key. The process is proportion to the number of...

>enhancement
:logstash

source: https://github.com/elastic/cloud-on-k8s/pull/6436#pullrequestreview-1349890633 During the e2e test, when more than 1 monitoring cluster referenced in `metrics`, the status states that all the associations are `Established`. This configuration is not supposed to...

>test
:logstash

Currently, the queue size is the sum of the event size per tail page and head page. Some pages may have 100kb event, some may have 30mb. However, pages are...

bug
persistent queues
status:confirmed

Version: 8.11.0 Logstash crashes when the pipeline id contain non-ASCII characters for example ``` input { http {} } filter { sleep { time => 1 id => "λ" }...

bug
int-shortlist
status:confirmed
Team:Logstash

[http.rb](https://github.com/ruby/net-http/blob/v0.2.2/lib/net/http.rb#L1225) use RUBY_PLATFORM to decide the source of username/ password of the proxy. RUBY_PLATFORM of jruby is `java`. It blocks jruby user from using proxy with env variable ruby --version...

The manticore adapter for the es-ruby client doesn’t implement `__close_connections`. Resources are unreleased when restarting a pipeline. Moreover, the plugin recreates the client when Elasticsearch is serverless, which further holds...

bug
status:approved

Currently, the database update scheduler runs every day at Logstash startup time. If client has thousand of instances running geoip filter at the same time, it could be a burden...

discuss

Users may not know their database license type. This commit adds supported commercial database name and CC database name

docs
status:needs-review

Since 7.14, geoip check database updates every 24h since the server start time. The suggestion is to add `interval` option to allow users to change the checking interval in hours....

enhancement
int-shortlist
papercut
status:confirmed

Currently, every [write](https://github.com/elastic/logstash/blob/8.13/logstash-core/src/main/java/org/logstash/ackedqueue/Queue.java#L420) and read [action](https://github.com/elastic/logstash/blob/8.13/logstash-core/src/main/java/org/logstash/ackedqueue/Queue.java#L617) shares the same lock. Write action always write to the head page, while the read action could read from the tail or head page...

performance improvements
persistent queues
Team:Logstash