Security Command Center is the security and risk database for Google Cloud. Security Command Center includes a risk dashboard and analytics system for surfacing, understanding, and remediating Google Cloud security and data risks across an organization.
Google Cloud Armor is integrated automatically with Security Command Center and exports two findings to the Security Command Center dashboard: Allowed Traffic Spike and Increasing Deny Ratio. This guide describes the findings and how to interpret them.
If you do not already have Google Cloud Armor enabled in Security Command Center, see Configuring Security Command Center. You see findings in Security Command Center only for projects that have Security Command Center enabled at the organization level.
Allowed Traffic Spike finding
Allowed traffic consists of well-formed HTTP(S) requests that are destined to reach your backend services after a Google Cloud Armor security policy is enforced.
The Allowed Traffic Spike finding notifies you of a spike in allowed traffic on a per-backend-service basis. A finding is generated when there is a sudden increase in the allowed number of requests per second (RPS) compared to the normal volume observed in recent history. The RPS that constituted the spike and the RPS of the recent history are provided as part of the finding.
Use case: Potential L7 attacks
Distributed denial-of-service (DDoS) attacks occur when attackers send large volumes of requests to overload a target service. Layer 7 DDoS attack traffic typically presents a spike in the number of requests per second.
An Allowed Traffic Spike finding identifies the backend service to which the RPS spike is directed and provides the traffic characteristics that caused Google Cloud Armor to classify it as an RPS spike. Use this information to determine the following:
- Whether a potential layer 7 DDoS attack is underway.
- The service that is being targeted.
- The actions that you can take to mitigate the potential attack.
The following is a screenshot of a sample Allowed Traffic Spike finding on the Security Command Center dashboard.
Google Cloud calculates the values Long_Term_Allowed_RPS and Short_Term_Allowed_RPS based on Google Cloud Armor historical information.
Increasing Deny Ratio finding
The Increasing Deny Ratio finding notifies you that there is an increase in the ratio of traffic that Google Cloud Armor blocks because of a user-configured rule in a security policy. Although the denial is expected and does not affect the backend service, this finding helps alert you to increases in unwanted and potentially malicious traffic targeting your applications. The RPS of the denied traffic and the total incoming traffic are provided as part of the finding.
Use case: Mitigating L7 attacks
An Increasing Deny Ratio finding enables you to see both the impact of successful mitigations and significant changes in the behavior of malicious clients. The finding identifies the backend to which the denied traffic was directed and provides the traffic characteristics that caused Google Cloud Armor to raise the finding. Use this information to evaluate whether the denied traffic must be studied in detail to further strengthen your mitigations.
The following is a screenshot of a sample Increasing Deny Ratio finding on the Security Command Center dashboard.
Google Cloud calculates the values Long_Term_Denied_RPS and Long_Term_Incoming_RPS based on Google Cloud Armor historical information.
Google Cloud Armor Adaptive Protection
Adaptive Protection sends telemetry to the Security Command Center. For more information about Adaptive Protection findings, see Monitoring, alerting, and logging in the Adaptive Protection overview.
After traffic returns to normal
Security Command Center findings are notifications that a particular behavior was observed at a point in time. No notification is sent when the behavior clears.
There might be updates to existing findings if the current traffic characteristics increase substantially in comparison to existing characteristics. If there is no follow-up finding, then either the behavior cleared or the traffic volume did not increase (allow or deny) substantially after the initial finding was generated.