Issue #, if available:Prometheus 3.0, released 11/14, removes the ability to use --enable-feature=agent --storage.agent.path="/opt/prometheus/data-agent flag which is implemented in the prometheus.service file by install-prometheus.sh (executed on HP controller node).
This leads to an error when the latest prometheus 3.0.0 version is pulled, with --enable-feature=agent path not recognized.
2 Alternatives explored, going with the lower risk option 2 until option 1 can be better understood.
1/ continuing to pull latest version and replacing with --storage.tsdb.path="/opt/prometheus/data" confirmed to work and successfully install prometheus correctly, however removing the remote-agent leads to risk of disk space filling up on controller. Until the risks of this new behavior are better understood, this option leads to too much risk.
2/ Pin the prometheus version to 2.55.1, which is known to work at x1000 node cluster scale.
Both options have been validated, but 2 is less risky.
Description of changes:
pin prometheus version to 2.55.1 instead of pulling latest.
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.
Issue #, if available: Prometheus 3.0, released 11/14, removes the ability to use
--enable-feature=agent --storage.agent.path="/opt/prometheus/data-agent
flag which is implemented in theprometheus.service
file by install-prometheus.sh (executed on HP controller node).This leads to an error when the latest prometheus 3.0.0 version is pulled, with
--enable-feature=agent
path not recognized.2 Alternatives explored, going with the lower risk option 2 until option 1 can be better understood.
1/ continuing to pull latest version and replacing with
--storage.tsdb.path="/opt/prometheus/data"
confirmed to work and successfully install prometheus correctly, however removing the remote-agent leads to risk of disk space filling up on controller. Until the risks of this new behavior are better understood, this option leads to too much risk.2/ Pin the prometheus version to 2.55.1, which is known to work at x1000 node cluster scale.
Both options have been validated, but 2 is less risky.
Description of changes: pin prometheus version to 2.55.1 instead of pulling latest.
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.