make a copy of the spreadsheet (you can also modify the copy as needed, for example: remove other product domain columns)
filter the metrics by product domain column (value should be 1) to only see metrics that were sent by your product domain
filter the metrics by column NumOfDomains to only see metrics that were sent by only 1..10 product domains (because some metrics are sent by our building blocks and that will involve almost all product domains -> backend-infra team will analyze those metrics)
add another column to list the link to dashboards/alerts that refer to your custom metrics
fill the column that was created in step (4), but please exclude links to dashboards/alerts that are auto-generated by Terraform, except if you change the dashboard/alert && use your custom metrics
remove unnecessary metrics or tags (need some code changes) -> metrics and tags that are not referenced anywhere in dashboard or alerts should be safe to remove

Please do so and share the spreadsheet to backend-infra team before June 30. Don't hesitate to contact us if you need any help

:bow:.

[Optional, will be discussed later with each team] For further checking:

check whether the total number of series make sense (see number of hosts, possible values of tags, etc)
check whether you need the metrics in the dashboards/alerts or not
check whether you need all the tags or not (you can see the detail at JSON dump (https://drive.google.com/open?id=1-x8sD5OSPRzt-aXw1OhLof2Z1zrQZL2k) or https://app.datadoghq.com/metric/summary?filter=[metricNameHere] -> and click the metric name on that page)

Note: regarding the high number of custom metrics on June 12 (more than 2 millions were sent), seems like it was caused by wrong aggregation by Datadog for the recent data.

Thank you!

Datadog Custom Metrics Details - Prod - June 2019