Bukhtawar Khan
Bukhtawar Khan
@saikaranam-amazon could you help elaborate on why that might be needed?
When we decommission the zone it is possible that traffic hasn't be DRAINED in which case it might take longer and calls getting timed out. The `GET` API can perform...
/cc : @ashking94 for help with review
@dblock I think GET with a body is considered as an ambiguous request and might lead to request smuggling. A payload within a GET/HEAD request message has no defined semantics....
/cc : @ankitkala
`NRTReplicationEngine` will unfortunately not be indexing into Lucene so LuceneChangesSnapshot will be empty https://github.com/opensearch-project/OpenSearch/blob/5444aac1684d26bfabbc17901b0a8584c700fd87/server/src/main/java/org/opensearch/index/engine/NRTReplicationEngine.java#L151-L161
Note that the current implementation although decouples translog from the engine, isn't a Pluggable modules yet. This change would just mean exposing a public method not necessarily pluggable since the...
cc-ing folks for comments @reta @dblock @andrross @nknize @sruti1312 @getsaurabh02
Agree with @dblock on the QoS, which is also the eventual goal once we are able to add more granular instrumentation on the latency breakdown across various layers and N/W...
Today there is no differentiation in the shard balancer algorithm in terms of placing primary and replica copies, both have equal weights. @hydrogen666 yes you are correct we would most...