Use API endpoint for series submission (DataDog#782)

* Use the API endpoint for timeseries submission of RDS enhanced metrics instead of the web endpoint * same for VPC Flow Logs metrics We in #fabric are trying to keep timeseries traffic out of the web endpoint, for better isolation and perforance tunning. This change aligns with our best practices, see https://docs.datadoghq.com/api/latest/metrics/#submit-metrics.

Author: simnyc

aws/rds_enhanced_monitoring/lambda_function.py

@@ -342,7 +342,7 @@ def flush(self):
  creds = urlencode(datadog_keys)
  data = json.dumps(metrics_dict).encode("ascii")
  url = "%s?%s" % (
- datadog_keys.get("api_host", "https://app.%s/api/v1/series" % DD_SITE),
+ datadog_keys.get("api_host", "https://api.%s/api/v1/series" % DD_SITE),
  creds,
  )
  req = Request(url, data, {"Content-Type": "application/json"})

aws/vpc_flow_log_monitoring/lambda_function.py

@@ -417,7 +417,7 @@ def flush(self):
  creds = urlencode(datadog_keys)
  data = json.dumps(metrics_dict).encode("utf-8")
  url = "%s?%s" % (
- datadog_keys.get("api_host", "https://app.%s/api/v1/series" % DD_SITE),
+ datadog_keys.get("api_host", "https://api.%s/api/v1/series" % DD_SITE),
  creds,
  )
  req = Request(url, data, {"Content-Type": "application/json"})