In TPCF/EAR versions prior to 10.4, the otel-collector runs as a monit process on the VMs in the CF deployment.
Process 'otel-collector' running
In 10.4 however, this has been replaced with the opamp-supervisor:
Process 'opamp-supervisor' running
TPCF/EAR 10.4
Looking at the config files for the opamp supervisor under /var/vcap/jobs/opamp-supervisor/config/opamp-supervisor.yml shows that it points to the otel-collector config/executable:
agent:
# Path to the collector binary that supervisor will manage
executable: /var/vcap/packages/otel-collector/otel-collector
# Using config_files so the supervisor can merge its auto-generated
# OpAMP extension configuration into the collector config.
config_files:
- $OWN_TELEMETRY_CONFIG
- $REMOTE_CONFIG
- /var/vcap/jobs/otel-collector/config/config.yml
- $OPAMP_EXTENSION_CONFIG
bootstrap_timeout: 30s
env:
GOMEMLIMIT: "614MiB"
Also, running ps -ef | grep collector shows the otel-collector executable running with relation to opamp-supervisor.
root 176400 62827 4 May28 ? 1-18:19:24 /var/vcap/packages/otel-collector/otel-collector --config /var/vcap/jobs/opamp-supervisor/store/effective.yaml --feature-gates service.AllowNoPipelines
So the opamp-supervisor is running and managing the otel collector in 10.4.2. More info on the opamp-supervisor can be found here:
The logs for the otel-collector itself can now be found in /var/vcap/jobs/opamp-supervisor/store/agent.log on each VM. However, there is currently no logrotation implemented on this file. A fix for this is in progress, and more details can be found here:
https://github.com/open-telemetry/opentelemetry-collector-contrib/issues/50376
https://github.com/open-telemetry/opentelemetry-collector-contrib/pull/50377