Investigate a deployment
Connect GitHub, GitLab or a deployment provider, then Sentry, Fly.io or an incident source for the same service. A regression needs a deployment, a signal increase and sufficient earlier history.
INTYGA Intelligence ยท Integrations
21 connectors, including the built-in INTYGA source. Choose a signal you need, check its coverage and connect it to the service it describes.
Reference implementations, checked with automated tests and representative provider payloads. Live-account validation is still needed for your permissions, provider plan and delivery configuration.
Bring changes and deployments together with the signals from the service they affect.
Pull requests, workflow runs and deployments; selected repository rule suites for merge-policy checks.
CI health and merge-policy monitoring need explicit targets. Sensitive paths create review observations.
Merge requests and deployments for code-change observations and deployment correlation.
GitHub-specific CI health and merge-policy checks do not automatically apply to GitLab.
Production deployment events for regression correlation.
Connect an error or incident source for the other side of the correlation.
Production deployment events for regression correlation.
Deployment events alone do not establish a regression.
Pages deployments for regression correlation.
This is Pages coverage, not a general Cloudflare security or Workers integration.
Use the alerts you already configure in your cloud provider. Collection does not scan every resource in your account.
CloudWatch alarm state, failed Backup jobs, AWS Health issues and GuardDuty findings.
Uses an API Destination with a shared secret. CloudTrail, Security Hub and cost analysis are not included.
Native Monitoring incidents, including open and closed states.
Uses a webhook notification channel with Basic authentication; not a Pub/Sub subscription.
Alerts Management alerts for a configured Azure subscription.
Uses an application credential. Supports public Azure; acknowledgement alone is not recovery.
Machine crashes and out-of-memory events for deployment regression correlation.
Not a complete inventory or resource-usage monitor.
Keep your existing monitoring tools. Intelligence uses their events as evidence or imports their alert conditions.
Error and issue events for deployment regression correlation.
Does not replace Sentry's issue investigation or tracing.
Monitor alerts for deployment regression correlation.
An alert is evidence from Datadog, not an independent root-cause determination.
Triggered incident events for deployment regression correlation.
Does not manage escalation policies or on-call schedules.
Triggered alerts for deployment regression correlation.
For existing Opsgenie setups; check the provider's current service availability before choosing it.
Grouped firing and resolved alerts, tracked separately per alert.
Uses the default webhook format with Basic authentication.
Grouped firing and resolved alerts, tracked separately per alert.
Uses webhook notifications with Basic authentication; this does not scrape Prometheus.
Incident status from the Incidents API, including explicit resolution.
Uses a team API token; response bodies and screenshots are not imported by this adapter.
PromQL query results and health checks for known INTYGA metrics.
Other queries can be collected, but arbitrary metrics are not automatically interpreted.
Investigate authorization patterns and provider-reported conditions beyond the deployment pipeline.
Authorization events for refusal-burst detection and deployment correlation.
The built-in source sees only activity routed through INTYGA.
Alert metadata and status from Workspace Alert Center.
Requires service-account delegation. Does not read Gmail messages or Drive document content.
Failed and paid payouts; created and closed disputes, using signed snapshot events.
Individual payment declines do not create findings. A closed dispute does not mean it was won.
Connect a source you control when an existing integration does not cover the signal you need.
Signed events for deployments, errors, CI outcomes, job completions and service-level aggregates.
Your service must emit the documented event format. Job and service-level detectors need configured targets.
Connect GitHub, GitLab or a deployment provider, then Sentry, Fly.io or an incident source for the same service. A regression needs a deployment, a signal increase and sufficient earlier history.
Start with an AWS, Google Cloud, Azure, Grafana, Alertmanager or Better Stack alert you already use. Confirm both its active state and its recovery reach Intelligence.
Emit signed completion events through the custom webhook. Configure the job's expected interval and grace period so a missed completion has a defined meaning.
Each connector's form explains its credential and delivery requirements. Webhook sources need an endpoint the provider can reach. Polling sources need API access and a running scheduler. Creating a connector leaves it switched off until you approve activation.
Native provider alerts are separate from deployment regression signals. Adding a new provider does not automatically make every detector understand its events.