Rerouting suspicious web traffic

1. A method comprising, by a logical entity in a network domain associated with an online application:

receiving a first request message associated with the online application from a client device, wherein the online application is provided by one or more application servers within the network domain, wherein the online application manages data corresponding to one or more users, and wherein the first request message is destined to one of the one or more application servers;

determining that a first suspicion score associated with the first request message exceeds a pre-determined threshold;

inserting into the first request message, in response to the determination, an indication configured to instruct the one of the one or more application servers to forward the first request message to a sandbox environment after processing the first request message, wherein the sandbox environment is configured to mimic the one or more application servers, and wherein the sandbox environment is configured to not update the data managed by the online application;

determining first routing policies based at least on the first suspicion score, the first routing policies indicating that the indication-inserted first request message is to be routed to the sandbox environment through the one of the one or more application servers as a new destination for the indication-inserted first request message; and

forwarding the indication-inserted first request message to a first computing device in the network domain based on the first routing policies.

2. The method of claim 1 , wherein the logical entity executes on one or more of the application servers.

3. The method of claim 1 , wherein the logical entity executes on a network device in the network domain, wherein the network device is a firewall, a gateway, or a traffic load-balancer.

4. The method of claim 1 , wherein the first suspicion score is calculated based on one or more features associated with the first request message, the one or more features comprising a source IP address associated with the client device, a geolocation information associated with the source IP address, traffic histories associated with the client device, or a reaction time corresponding to the first request message in response to a previous message from the one or more application servers to the client device.

5. The method of claim 4 , wherein the first suspicion score is calculated by a security engine.

6. The method of claim 5 , wherein the security engine executes on a device that the logical entity executes on.

7. The method of claim 5 , wherein the security engine executes on a different device in the network domain from a device that the logical entity executes on.

8. The method of claim 5 , wherein the security engine is executed by a third-party device and is accessible through a call to an application programming interface provided by the third-party device.

9. The method of claim 5 , wherein the first suspicion score is calculated with a trained machine-learning model.

10. The method of claim 1 , wherein determining that the first suspicion score associated with the first request message exceeds the pre-determined threshold comprises receiving a confirmation on the first suspicion score from a security administrator.

11. The method of claim 1 , wherein the first request message comprises state information associated with previous interactions between the one or more application servers and the client device.

12. The method of claim 1 , further comprising:

receiving a second request message associated with the online application from the client device;

determining that a second suspicion score associated with the second request message is lower than the pre-determined threshold;

determining second routing policies based on the second suspicion score, the second routing policies indicating that the second request message is to be routed to one of the one or more application servers; and

forwarding the second request message to a second computing device in the network domain based on the second routing policies.

13. The method of claim 1 , wherein the first computing device is one of the one or more application servers.

14. The method of claim 13 , wherein the one of the one or more application servers processes the first request message before forwarding the first request message to the sandbox environment.

15. The method of claim 1 , wherein the first computing device belongs to the sandbox environment.

16. The method of claim 1 , wherein the indication is a tag inserted into a header of the first request message.

17. The method of claim 1 , wherein the sandbox environment comprises one or more computing devices.

18. The method of claim 1 , wherein the sandbox environment comprises one or more virtual machines.

19. One or more computer-readable non-transitory storage media embodying software that is operable when executed, by a logical entity in a network domain associated with an online application, to:

receive a first request message associated with the online application from a client device, wherein the online application is provided by one or more application servers within the network domain, wherein the online application manages data corresponding to one or more users, and wherein the first request message is destined to one of the one or more application servers;

determine that a first suspicion score associated with the first request message exceeds a pre-determined threshold;

insert into the first request message, in response to the determination, an indication configured to instruct the one of the one or more application servers to forward the first request message to a sandbox environment after processing the first request message, wherein the sandbox environment is configured to mimic the one or more application servers, and wherein the sandbox environment is configured to not update the data managed by the online application;

determine first routing policies based at least on the first suspicion score, the first routing policies indicating that the indication-inserted first request message is to be forwarded to the sandbox environment as a new destination for the indication-inserted first request message; and

forward the indication-inserted first request message to a first computing device in the network domain based on the first routing policies.

20. A system comprising:

one or more processors; and

one or more computer-readable non-transitory storage media coupled to one or more of the processors and comprising instructions operable when executed by one or more of the processors to cause a logical entity in a network domain associated with an online application executing on the system to:

receive a first request message associated with the online application from a client device, wherein the online application is provided by one or more application servers within the network domain, wherein the online application manages data corresponding to one or more users, and wherein the first request message is destined to one of the one or more application servers;

determine that a first suspicion score associated with the first request message exceeds a pre-determined threshold;

insert into the first request message, in response to the determination, an indication configured to instruct the one of the one or more application servers to forward the first request message to a sandbox environment after processing the first request message, wherein the sandbox environment is configured to mimic the one or more application servers, and wherein the sandbox environment is configured to not update the data managed by the online application;

determine first routing policies based at least on the first suspicion score, the first routing policies indicating that the indication-inserted first request message is to be forwarded to the sandbox environment as a new destination for the indication-inserted first request message; and

forward the indication-inserted first request message to a first computing device in the network domain based on the first routing policies.