Upgrade to Observe Agent v1.0.0¶
Observe Agent v1.0.0 includes the following breaking changes. Before upgrading the previous versions of Observe Agent to v1.0.0 (Linux, Windows, Ansible), please check the following breaking changes.
Per-process metrics are not collected by default¶
Since per-process metrics could produce a lot of metrics when you have many processes running, starting from v1.0.0, the Observe Agent doesn’t collect per-process metrics by default. If you need to collect per-process metrics, use the following config block to re-configure the Observe Agent v1.0.0 (Linux, Windows, Ansible).
self_monitoring:
enabled: true
host_monitoring:
enabled: true
logs:
enabled: true
metrics:
host:
enabled: true
process:
enabled: true
init-config
¶
A breaking change is being made to the init-config
command, specifically to the format of the flags that define the configuration values to be written to. Previously the flags for a given config field were of the format host_monitoring.logs.enabled
. Moving forward, ::
will be used as the delimiter for nested key paths. The new flags will be of the format host_monitoring::logs::enabled
. This should only affect new installations when using init-config
to generate the initial config. Previously generated configs will continue to be valid without any changes.
self_monitoring
Connection Type¶
Part of the initial Observe Agent configuration that collected metrics and logs from the agent itself was previously enabled by default. Going forward, the self_monitoring
connection type has been added and the config for that data has been migrated under this connection type. New installations of the Observe Agent will set this field to true by default in the observe-agent.yaml
but for upgrading from existing versions, if your current observe-agent.yaml
config file doesn’t configure the self_monitoring connection type then the Observe Agent will stop collecting this metrics and logs data. To enable, you can add the following section to your observe-agent.yaml
self_monitoring:
enabled: true
collection_interval
increased to 60 seconds¶
collection_interval
determines how often host metrics are being collected. Observe Agent v1.0.0 increases the collection_interval
to 60 seconds from 20 seconds. If you want to reduce the collection_interval
in order to collect host metrics more often, you can add the following section to your observe-agent.yaml
otel_config_overrides:
receivers:
hostmetrics/host-monitoring-host:
collection_interval: 20s