Rick Boyd
Rick Boyd
## Background Recently @salvatore-campagna and others added multiple aggregations-centric workloads to our nightly benchmarks at https://elasticsearch-benchmarks.elastic.co/. One in particular has been failing frequently and was recently removed; the `aggs` challenge...
## Background Recently @salvatore-campagna and others added multiple aggregations-centric workloads to our nightly benchmarks at https://elasticsearch-benchmarks.elastic.co/. One in particular has been failing frequently and was recently removed; the `aggs` challenge...
Recently to investigate tail latencies for our `knn` search in Elasticsearch, we've added the capability to profile the path of execution in a search transaction, yielding traces that look like...
It would be nice if track generation could optionally include document _id when reading corpora from an existing index. Discuss issue: https://discuss.elastic.co/t/set-custom-document-ids-on-bulk-insert/258455/
We should have a native operation in Rally for adding ingest pipelines. example: ```json { "name": "insert-pipelines", "operation-type": "create-pipeline", "pipelines": "pipelines/*" } ``` Note `pipelines` above as a file glob,...
We should add an `apm` telemetry device to take advantage of the `tracing.apm` configurations available in Elasticsearch, to send trace information to a cluster. Bonus if it can somehow default...
**Rally version** (get with `esrally --version`): 2.6.1.dev0 **Description of the problem including expected versus actual behavior**: When pointing Rally to use `--target-hosts` for sockets that are already occupied, it logs...
This problem is in current master, 2.3.1dev0. If I execute ``` esrally install --distribution-version 7.16.0 --car 4gheap,trial-license --http-port=39200 --master-nodes="rally-node-0" --seed-hosts="127.0.0.1:39300" esrally start --installation-id --race-id 1234 esrally race --distribution-version 7.16.1 --track...
As of #1190, `scroll-search` is its own operation, and as such we should remove scroll support (aka the usage of the "pages" parameter) for search operations.
Because `wait-for-recovery` can be a long-running operation, it would be nice and probably not too difficult to implement the `percent_completed` feedback to the console. This used to be implemented but...