Reputation: 2918
we have the datadog agent manually setup on our Kubernetes Cluster which helped us forwarding Logs and APM from our application containers into DataDog.
Recently we noticed the biggest item in our bill is "On-Demand Infrastructure Container Hours". Which seems to be related to this view:
We really do not use that at all, and maybe we integrated incorrectly but right now we get no value out of it.
Is there a way of getting the Logs and APM from containers, but avoiding the big bill on this feature?
In the table on the manual installation it seems metrics is non-optional for the Agent. And I'm guessing this is what "metrics" is.
Upvotes: 2
Views: 4206
Reputation: 111
As per Datadog docs, there are following items in the Infrastructure moniroting:
What you're showing in the screenshot is the list of Hosts.
But the billing item "On-Demand Infrastructure Container Hours" refers to Containers.
To disable infra monitoring for some containers, you'd need to exclude them via the DD_CONTAINER_EXCLUDE env variable. See the docs
E.g. to exclude ALL containers:
DD_CONTAINER_EXCLUDE = "name:.*"
Bonus: To also disable Hosts monitoring there's a less obvious configuration that needs to be applied (see docs) - add the following env variables to disable certain payloads:
env:
- name: DD_ENABLE_PAYLOADS_EVENTS
value: "false"
- name: DD_ENABLE_PAYLOADS_SERIES
value: "false"
- name: DD_ENABLE_PAYLOADS_SERVICE_CHECKS
value: "false"
- name: DD_ENABLE_PAYLOADS_SKETCHES
value: "false"
However as per docs, it may affect the logs collecting, so please verify if it affects your particular case.
Upvotes: 2