Add _reload_search_analyzers endpoint to HLRC #43733
Conversation
This change adds the new endpoint that allows reloading of search analyzers to the high-level java rest client. Relates to elastic#43313
cbuescher
added
>enhancement
:Search Relevance/Analysis
|
Pinging @elastic/es-search |
|
Pinging @elastic/es-core-features |
|
@elasticmachine update branch |
|
@elasticmachine update branch |
| -------------------------------------------------- | ||
| include-tagged::{doc-tests-file}[{api}-response] | ||
| -------------------------------------------------- | ||
| <1> Shard statistics including total number of shards the request was executed on |
There was a problem hiding this comment.
This is not entirely related to this pr but I find it confusing that we return a Shards object in the response even though the action is executed at the node level. I don't have a good way to fix this but we should at least document here that the statistics are per node/index and not per shards ?
There was a problem hiding this comment.
I agree, this is not ideal. I'll look into the options of either documenting it better, change the naming of the object only for the java api level or whether maybe it makes sense change or even leave out the whole "shard: { ... }". That would have to be done further down the stack though.
There was a problem hiding this comment.
I looked into the options, and it would be possible but tricky to remove the "shard: { ... }" element from the REST response completely. TransportReloadAnalyzersAction would still be able to extend BroadcastResponse, but we could overwrite "toXContent" without outputing broadcast shards header. Then again, potential failure exceptions are currently also written inside the "shards" header, replacing that toXContent rendering and parsing might turn out to be a bigger change than intended.
Changing the accessor or object name would also involve either not being able to extend BroadcastResponse on the HLRC client side (this would mean more custom parsing code, other getters etc... for failures), so I'll push an update to the docs for now. If you want to discuss the other options again let me know.
Currently the repsonse of the "_reload_search_analyzer" endpoint contains the index names and nodeIds of indices were analyzers reloading was triggered. This change add the names of the search-time analyzers that were reloaded. Closes elastic#43804
|
@jimczi as this PR depends on #43813 and CI is on strike right now I cherry-picked that PR in |
|
@elasticmachine test this please |
|
@elasticmachine run elasticsearch-ci/1 |
|
@elasticmachine run elasticsearch-ci/2 |
|
@elasticmachine run elasticsearch-ci/2 |
This change adds the new endpoint that allows reloading of search analyzers to the high-level java rest client. Relates to #43313
This change adds the new endpoint that allows reloading of search analyzers to
the high-level java rest client.
Relates to #43313