Datadog integration

DevFlow can push every check result to Datadog as both a metric and (for failures) an event.

install

In Datadog, create an API key (Org Settings → API Keys). In DevFlow:

devflow integrations datadog configure   --api-key dd_abcdef   --site us5.datadoghq.com

--site defaults to datadoghq.com. EU and Gov sites supported.

what's emitted

For every check, two metrics:

• devflow.check.latency — gauge, ms, tagged with monitor, region, env, service.
• devflow.check.success — gauge, 0 or 1.

For every failed check, an event:

• Title — DevFlow: <monitor> failed in <region>.
• Tags — same as metrics.
• Source type name — devflow.

For every incident open/ack/resolve, an event with source_type:devflow and a devflow.incident.lifecycle tag.

what to do with it

Most teams add a Datadog dashboard widget for devflow.check.latency per service-tag, alongside their service's own latency metrics. The two side-by-side are the most useful artefact — your service vs DevFlow's external view.

limits

We respect Datadog's rate limits on /v1/series and back off on 429. For very-high-frequency monitors (sub-10s, Scale tier), we coalesce into 10s buckets before posting.

removing

Revoke the API key in Datadog. DevFlow surfaces the failure on the next delivery; the integration auto-disables after 1h of failed deliveries.