Skip to main content
This integration receives Amazon EventBridge events, either pushed directly by an API destination or relayed by an SNS topic. CloudWatch alarm, GuardDuty, Security Hub, AWS Health, Amazon Inspector, AWS Config and DevOps Guru events merge per alarm or finding and recover automatically when the source recovers. Every other event creates one alert.

In Flashduty On-call

You can get the integration push URL in either of the following ways.

Use a dedicated integration

Choose this option if all alerts go to one channel. It’s the simpler setup.
  1. In the Flashduty console, select Channel and open a channel
  2. Select Configuration → Integrations → Private integration, then click Add integration
  3. Choose AWS EventBridge integration and click Save to generate a card
  4. Open the generated integration card and copy the Push URL

Use a shared integration

Choose this option if you need to route alerts to different channels based on the alert payload.
  1. In the Flashduty console, go to Integration Center → Alert Events
  2. Select AWS EventBridge integration:
    • Integration Name: Define a name for this integration
  3. Configure the default route and select a channel. You can add more rules under Routing after creation
  4. Click Save and copy the generated Push URL

Configure in AWS EventBridge

1

Create an API destination that points to Flashduty in Amazon EventBridge

2

Create an event rule in Amazon EventBridge

Where: Amazon EventBridge › Rules › Create rule › Advanced BuilderUse an event pattern in the rule, for example:
Do not filter out recovery states in the rule, or alerts cannot recover automatically. For example, matching only "state": {"value": ["ALARM"]}, "newEvaluationResult": {"complianceType": ["NON_COMPLIANT"]}, Security Hub "findings": {"Workflow": {"Status": ["NEW"]}} or Inspector "status": ["ACTIVE"] drops the recovery events.
AWS Health sends events to the Region where they occur; events of global services arrive in us-east-1. Create the rule in every Region you care about.Select the API destination as the target of an EventBridge rule
3

Send a test CloudWatch alarm event and confirm

Where: A terminal that can reach the Flashduty push URLThe rule only forwards events that AWS really emits. To verify right away, POST a test event shaped like the real one to the push URL:
Send it again with state.value set to OK to confirm that the alert recovers automatically. A few seconds later, go back to Flashduty. The integration shows a latest event time, which means the setup works.

Alert key and recovery

  • Later events of the same source (repeated GuardDuty findings, Security Hub updates, AWS Health progress) merge into the original alert
  • When one Security Hub event carries several findings, each finding creates its own alert
  • An event missing the Alert Key field listed above is rejected (HTTP 400)
  • GuardDuty sends no event when a finding is archived, so close the alert manually

Status and severity

EventBridge events have no common severity field, so every event triggers as Warning, as it always has. The source’s own severity is kept in a label, and an alert pipeline can rewrite the severity based on it: For CloudWatch alarms, set the severity with flashduty_severity: critical (critical, warning or info) in the Alarm description, the same way as in the AWS CloudWatch integration.

Title and labels

  • The alert title is source::detail-type, for example aws.guardduty::GuardDuty Finding
  • Every event carries the labels source, region, account, check (detail-type), detail (the event detail as JSON) and resources
  • The summary label is a readable summary: the finding or alarm title for sources with dedicated handling; for other events, in order, CloudTrail eventSource eventName, RDS SourceIdentifier Message, ECS group stoppedReason, or EC2 instance-id state
  • Sources with dedicated handling also add source fields such as finding_id, finding_arn, event_arn, config_rule_name and insight_id; resource is the affected resource (DevOps Guru insights have none)

Troubleshooting

  • API destination invocations fail: Make sure the endpoint is the full push URL including integration_key and HTTP method is POST
  • SNS subscription stays in Pending confirmation: Make sure the subscription points to this integration’s push URL, then run Request confirmation in the SNS console
  • Alert does not recover: Check whether the rule’s event pattern filters out recovery states, and whether the target has an input transformer
  • One finding creates several alerts: Make sure the event comes from a source in the table above and its detail-type is not rewritten