Here are some ways of getting your metrics data into Datadog:
- We provide an SDK for people interested in writing new integrations.
- A common method is via DogStatsD, the StatsD server bundled with the Datadog Agent. To learn more read our guide along with this great article.
- Submit metrics directly to Datadog's HTTP API
- Another method to consider is using dogstatsd and the shell
- Multiple community libraries are listed here (Go, Python, PHP, Yammer, Node.js, etc.)
Another option is to extend or fork any of our integrations that run on your localhost, you may learn more about this here.
Be aware that pushing additional metrics to Datadog may impact your billing, please review this article to learn more.
If you're submitting metrics directly to the Datadog API sans DogStatsD, expect:
- 64 bits for the timestamp
- 64 bits for the value
- 20bytes for the metric names
- 50 bytes for the time-series
Full payload being ~100 bytes.
However, if you're using the DogStatsD API a typical payload is very small over the wire as compression is applied.