I am working through an open support ticket, and one of the requests from the engineer was to pull diagnostics (GitHub - elastic/support-diagnostics: Support diagnostics utility for elasticsearch and logstash) for my cluster. It got me thinking this might be a worthwhile enhancement:
Our cluster uses the official Elasticsearch docker image. I know I could just use the "remote" option on diagnostics, but that disables collection of logs/configs. I'd like to avoid that if possible and collect as much as I can.
I am wondering if it would make sense to include a build of the diagnostics tool (and dependencies) inside the official ES Docker image? I found it a bit awkward to mount the diagnostics into the container to run inside, and when I do so some of the tool dependencies are missing from the alpine container (e.g. jps, jstack, ss/netstat).
Maybe also allow the diagnostic tool to stream the output to stdout, so I can do something like "docker exec -it Elasticsearch diagnostics.sh --host localhost --stdout" to be able to pipe it into something outside the running container?