Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
Server-side request forgery tricks an application into making HTTP requests of the attacker’s choosing, reaching internal services and cloud metadata that are otherwise unreachable.

What Is Server-Side Request Forgery?
Server-side request forgery turns an application into a proxy for the attacker. A feature that fetches a URL, perhaps to import an image, render a preview, call a webhook, or convert a document, accepts a destination from the user. Because the request originates from your server, it arrives at its destination with your server’s network position and often with its credentials attached. Anything the application can reach, the attacker can now reach.
The most valuable targets are the ones that were never designed to be exposed, because their owners assumed network isolation was sufficient:
http://169.254.169.254/ returns instance information and, in older configurations, temporary credentials for the attached role.Blind variants, where the response is never shown to the attacker, are still useful: response timing and error differences map the internal network, and a request to an attacker-controlled host confirms the vulnerability and may carry data in the URL.
The case that made boards pay attention was disclosed on 19 July 2019. An attacker exploited a server-side request forgery in a web application firewall configuration at Capital One, used it to query the EC2 instance metadata service, and retrieved temporary credentials for the role attached to the instance. Those credentials permitted listing and reading S3 buckets, from which data relating to approximately 100 million US and 6 million Canadian applicants and customers was taken. The Office of the Comptroller of the Currency imposed an $80 million civil penalty in August 2020, citing failures in risk assessment and remediation rather than in any single line of code.
The architectural lesson is that an application reachable from the internet held a cloud role with broad data access, and the metadata service version in use handed out credentials to any process that asked. IMDSv2, which requires a PUT request to obtain a session token and sets a hop limit that prevents proxied access, removes that specific path and should be enforced everywhere.
Teams usually respond by blocking internal addresses, and attackers have a long list of answers. A DNS name under the attacker’s control can resolve to a private address. Redirects move the request after validation has passed. Alternative encodings of an address, decimal or octal forms, and IPv6 representations defeat string comparison. The file://, gopher://, and dict:// schemes reach beyond HTTP. DNS rebinding resolves to a permitted address for the check and a private one for the connection, which defeats any design that validates and fetches separately.
OWASP listed SSRF as its own category, A10, in the 2021 Top 10 after a community survey ranked it unusually high relative to its measured incidence; the 2025 revision folded it into the broken access control category, which reflects what it is at root. Testing means supplying every URL-accepting parameter with an internal address, a collaborator domain, a redirect, and an alternative scheme, then watching for interaction rather than for a visible response. Our penetration test services include internal reachability testing from the application tier, which is what turns an SSRF finding into an accurate statement about risk.
No, though cloud environments raise the impact considerably because of the metadata service and the credentials it can expose. On-premises, the same vulnerability reaches internal web applications, unauthenticated databases, management interfaces, and printers. Any environment where services trust requests arriving from the internal network is exposed, and most internal networks are built on exactly that assumption.
Yes. Without seeing responses, an attacker can still map internal hosts and ports from timing and error differences, trigger state-changing requests against internal services that accept GET, and exfiltrate small amounts of data through DNS lookups. Several published critical findings were blind, and the inability to see responses often slows exploitation rather than preventing it.
Not on its own. Attackers use hostnames that resolve to private addresses, redirects that move the request after validation, alternative numeric encodings, non-HTTP schemes, and DNS rebinding where the name resolves differently between the check and the connection. Resolve once and connect to the validated address, and enforce the policy at an egress proxy rather than in application code.
It is the hardened version of the cloud instance metadata service. A client must first obtain a session token with a PUT request carrying a specific header, and the response hop limit prevents the token being relayed through a proxy or a simple forged GET. That combination defeats the straightforward SSRF path to instance credentials, which is why enforcing it across all instances is worthwhile.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.