AutoOps support

[smith]

Allow running AutoOps agent as described here: https://www.elastic.co/blog/elasticsearch-autoops-on-prem

The agent is another standalone collector that runs metricbeatreceiver, so it may be possible to run the necessary pipelines in the edot collector.

The settings env vars could probably point to a cloud cluster we share by default, but leaving out the secrets/api key because this is a private repo.

This is the command the installation UI in kibana gives you:

docker run -d \
-e AUTOOPS_OTEL_URL='https://otel-auto-ops.us-east-2.aws.svc.elastic.cloud' \
-e AUTOOPS_TOKEN='xxxxxxxxxx' \
-e AUTOOPS_ES_URL='https://elasticsearch:9200' \
-e ELASTIC_CLOUD_CONNECTED_MODE_API_KEY='xxx' \
-e AUTOOPS_ES_USERNAME='{{username}}' \
-e AUTOOPS_ES_PASSWORD='{{password}}' \
-e AUTOOPS_TEMP_RESOURCE_ID='d611905b-c59a-4237-9b45-691be2bd6f0c' \
--restart on-failure --name autoops-otel-agent \
docker.elastic.co/elastic-agent/elastic-otel-collector-wolfi:9.1.4 \
--config otel_samples/autoops_es.yml \
--set "exporters::otlphttp::sending_queue::batch::max_size=4194304" \
--set "exporters::otlphttp::sending_queue::batch::sizer=bytes"

(This doesn't work with the existing setup when not run as part of the docker compose setup. Using host.docker.internal instead of elasticsearch probably would.)