This commit introduces two new adaptive concurrency limiters in Vault, which should handle overloading of the server during periods of untenable request rate. The limiter adjusts the number of allowable in-flight requests based on latency measurements performed across the request duration. This approach allows us to reject entire requests prior to doing any work and prevents clients from exceeding server capacity. The limiters intentionally target two separate vectors that have been proven to lead to server over-utilization. - Back pressure from the storage backend, resulting in bufferbloat in the WAL system. (enterprise) - Back pressure from CPU over-utilization via PKI issue requests (specifically for RSA keys), resulting in failed heartbeats. Storage constraints can be accounted for by limiting logical requests according to their http.Method. We only limit requests with write-based methods, since these will result in storage Puts and exhibit the aforementioned bufferbloat. CPU constraints are accounted for using the same underlying library and technique; however, they require special treatment. The maximum number of concurrent pki/issue requests found in testing (again, specifically for RSA keys) is far lower than the minimum tolerable write request rate. Without separate limiting, we would artificially impose limits on tolerable request rates for non-PKI requests. To specifically target PKI issue requests, we add a new PathsSpecial field, called limited, allowing backends to specify a list of paths which should get special-case request limiting. For the sake of code cleanliness and future extensibility, we introduce the concept of a LimiterRegistry. The registry proposed in this PR has two entries, corresponding with the two vectors above. Each Limiter entry has its own corresponding maximum and minimum concurrency, allowing them to react to latency deviation independently and handle high volumes of requests to targeted bottlenecks (CPU and storage). In both cases, utilization will be effectively throttled before Vault reaches any degraded state. The resulting 503 - Service Unavailable is a retryable HTTP response code, which can be handled to gracefully retry and eventually succeed. Clients should handle this by retrying with jitter and exponential backoff. This is done within Vault's API, using the go-retryablehttp library. Limiter testing was performed via benchmarks of mixed workloads and across a deployment of agent pods with great success. |
||
|---|---|---|
| .github | ||
| .hooks | ||
| .release | ||
| api | ||
| audit | ||
| builtin | ||
| changelog | ||
| command | ||
| dependencies/2-25-21 | ||
| enos | ||
| helper | ||
| http | ||
| internal/observability/event | ||
| internalshared | ||
| limits | ||
| META.d | ||
| physical | ||
| plugins/database | ||
| scripts | ||
| sdk | ||
| serviceregistration | ||
| shamir | ||
| terraform | ||
| tools | ||
| ui | ||
| vault | ||
| version | ||
| website | ||
| .copywrite.hcl | ||
| .gitattributes | ||
| .gitignore | ||
| .go-version | ||
| .yamllint | ||
| buf.gen.yaml | ||
| buf.lock | ||
| buf.yaml | ||
| CHANGELOG-pre-v1.10.md | ||
| CHANGELOG-v0.md | ||
| CHANGELOG.md | ||
| CODEOWNERS | ||
| CONTRIBUTING.md | ||
| Dockerfile | ||
| go.mod | ||
| go.sum | ||
| HCPV_badge.png | ||
| LICENSE | ||
| main.go | ||
| main_test.go | ||
| make.bat | ||
| Makefile | ||
| README.md | ||
| scan.hcl | ||
Vault

Please note: We take Vault's security and our users' trust very seriously. If you believe you have found a security issue in Vault, please responsibly disclose by contacting us at security@hashicorp.com.
- Website: https://www.vaultproject.io
- Announcement list: Google Groups
- Discussion forum: Discuss
- Documentation: https://developer.hashicorp.com/vault/docs
- Tutorials: https://developer.hashicorp.com/vault/tutorials
- Certification Exam: https://developer.hashicorp.com/certifications/security-automation
Vault is a tool for securely accessing secrets. A secret is anything that you want to tightly control access to, such as API keys, passwords, certificates, and more. Vault provides a unified interface to any secret, while providing tight access control and recording a detailed audit log.
A modern system requires access to a multitude of secrets: database credentials, API keys for external services, credentials for service-oriented architecture communication, etc. Understanding who is accessing what secrets is already very difficult and platform-specific. Adding on key rolling, secure storage, and detailed audit logs is almost impossible without a custom solution. This is where Vault steps in.
The key features of Vault are:
-
Secure Secret Storage: Arbitrary key/value secrets can be stored in Vault. Vault encrypts these secrets prior to writing them to persistent storage, so gaining access to the raw storage isn't enough to access your secrets. Vault can write to disk, Consul, and more.
-
Dynamic Secrets: Vault can generate secrets on-demand for some systems, such as AWS or SQL databases. For example, when an application needs to access an S3 bucket, it asks Vault for credentials, and Vault will generate an AWS keypair with valid permissions on demand. After creating these dynamic secrets, Vault will also automatically revoke them after the lease is up.
-
Data Encryption: Vault can encrypt and decrypt data without storing it. This allows security teams to define encryption parameters and developers to store encrypted data in a location such as a SQL database without having to design their own encryption methods.
-
Leasing and Renewal: All secrets in Vault have a lease associated with them. At the end of the lease, Vault will automatically revoke that secret. Clients are able to renew leases via built-in renew APIs.
-
Revocation: Vault has built-in support for secret revocation. Vault can revoke not only single secrets, but a tree of secrets, for example, all secrets read by a specific user, or all secrets of a particular type. Revocation assists in key rolling as well as locking down systems in the case of an intrusion.
Documentation, Getting Started, and Certification Exams
Documentation is available on the Vault website.
If you're new to Vault and want to get started with security automation, please check out our Getting Started guides on HashiCorp's learning platform. There are also additional guides to continue your learning.
For examples of how to interact with Vault from inside your application in different programming languages, see the vault-examples repo. An out-of-the-box sample application is also available.
Show off your Vault knowledge by passing a certification exam. Visit the certification page for information about exams and find study materials on HashiCorp's learning platform.
Developing Vault
If you wish to work on Vault itself or any of its built-in systems, you'll first need Go installed on your machine.
For local dev first make sure Go is properly installed, including setting up a
GOPATH. Ensure that $GOPATH/bin is in
your path as some distributions bundle the old version of build tools. Next, clone this
repository. Vault uses Go Modules,
so it is recommended that you clone the repository outside of the GOPATH.
You can then download any required build tools by bootstrapping your environment:
$ make bootstrap
...
To compile a development version of Vault, run make or make dev. This will
put the Vault binary in the bin and $GOPATH/bin folders:
$ make dev
...
$ bin/vault
...
To compile a development version of Vault with the UI, run make static-dist dev-ui. This will
put the Vault binary in the bin and $GOPATH/bin folders:
$ make static-dist dev-ui
...
$ bin/vault
...
To run tests, type make test. Note: this requires Docker to be installed. If
this exits with exit status 0, then everything is working!
$ make test
...
If you're developing a specific package, you can run tests for just that
package by specifying the TEST variable. For example below, only
vault package tests will be run.
$ make test TEST=./vault
...
Importing Vault
This repository publishes two libraries that may be imported by other projects:
github.com/hashicorp/vault/api and github.com/hashicorp/vault/sdk.
Note that this repository also contains Vault (the product), and as with most Go
projects, Vault uses Go modules to manage its dependencies. The mechanism to do
that is the go.mod file. As it happens, the presence of that file
also makes it theoretically possible to import Vault as a dependency into other
projects. Some other projects have made a practice of doing so in order to take
advantage of testing tooling that was developed for testing Vault itself. This
is not, and has never been, a supported way to use the Vault project. We aren't
likely to fix bugs relating to failure to import github.com/hashicorp/vault
into your project.
See also the section "Docker-based tests" below.
Acceptance Tests
Vault has comprehensive acceptance tests covering most of the features of the secret and auth methods.
If you're working on a feature of a secret or auth method and want to verify it is functioning (and also hasn't broken anything else), we recommend running the acceptance tests.
Warning: The acceptance tests create/destroy/modify real resources, which may incur real costs in some cases. In the presence of a bug, it is technically possible that broken backends could leave dangling data behind. Therefore, please run the acceptance tests at your own risk. At the very least, we recommend running them in their own private account for whatever backend you're testing.
To run the acceptance tests, invoke make testacc:
$ make testacc TEST=./builtin/logical/consul
...
The TEST variable is required, and you should specify the folder where the
backend is. The TESTARGS variable is recommended to filter down to a specific
resource to test, since testing all of them at once can sometimes take a very
long time.
Acceptance tests typically require other environment variables to be set for things such as access keys. The test itself should error early and tell you what to set, so it is not documented here.
For more information on Vault Enterprise features, visit the Vault Enterprise site.
Docker-based Tests
We have created an experimental new testing mechanism inspired by NewTestCluster. An example of how to use it:
import (
"testing"
"github.com/hashicorp/vault/sdk/helper/testcluster/docker"
)
func Test_Something_With_Docker(t *testing.T) {
opts := &docker.DockerClusterOptions{
ImageRepo: "hashicorp/vault", // or "hashicorp/vault-enterprise"
ImageTag: "latest",
}
cluster := docker.NewTestDockerCluster(t, opts)
defer cluster.Cleanup()
client := cluster.Nodes()[0].APIClient()
_, err := client.Logical().Read("sys/storage/raft/configuration")
if err != nil {
t.Fatal(err)
}
}
Or for Enterprise:
import (
"testing"
"github.com/hashicorp/vault/sdk/helper/testcluster/docker"
)
func Test_Something_With_Docker(t *testing.T) {
opts := &docker.DockerClusterOptions{
ImageRepo: "hashicorp/vault-enterprise",
ImageTag: "latest",
VaultLicense: licenseString, // not a path, the actual license bytes
}
cluster := docker.NewTestDockerCluster(t, opts)
defer cluster.Cleanup()
}
Here is a more realistic example of how we use it in practice. DefaultOptions uses
hashicorp/vault:latest as the repo and tag, but it also looks at the environment
variable VAULT_BINARY. If populated, it will copy the local file referenced by
VAULT_BINARY into the container. This is useful when testing local changes.
Instead of setting the VaultLicense option, you can set the VAULT_LICENSE_CI environment variable, which is better than committing a license to version control.
Optionally you can set COMMIT_SHA, which will be appended to the image name we build as a debugging convenience.
func Test_Custom_Build_With_Docker(t *testing.T) {
opts := docker.DefaultOptions(t)
cluster := docker.NewTestDockerCluster(t, opts)
defer cluster.Cleanup()
}
There are a variety of helpers in the github.com/hashicorp/vault/sdk/helper/testcluster
package, e.g. these tests below will create a pair of 3-node clusters and link them using
PR or DR replication respectively, and fail if the replication state doesn't become healthy
before the passed context expires.
Again, as written, these depend on having a Vault Enterprise binary locally and the env var VAULT_BINARY set to point to it, as well as having VAULT_LICENSE_CI set.
func TestStandardPerfReplication_Docker(t *testing.T) {
opts := docker.DefaultOptions(t)
r, err := docker.NewReplicationSetDocker(t, opts)
if err != nil {
t.Fatal(err)
}
defer r.Cleanup()
ctx, cancel := context.WithTimeout(context.Background(), time.Minute)
defer cancel()
err = r.StandardPerfReplication(ctx)
if err != nil {
t.Fatal(err)
}
}
func TestStandardDRReplication_Docker(t *testing.T) {
opts := docker.DefaultOptions(t)
r, err := docker.NewReplicationSetDocker(t, opts)
if err != nil {
t.Fatal(err)
}
defer r.Cleanup()
ctx, cancel := context.WithTimeout(context.Background(), time.Minute)
defer cancel()
err = r.StandardDRReplication(ctx)
if err != nil {
t.Fatal(err)
}
}
Finally, here's an example of running an existing OSS docker test with a custom binary:
$ GOOS=linux make dev
$ VAULT_BINARY=$(pwd)/bin/vault go test -run 'TestRaft_Configuration_Docker' ./vault/external_tests/raft/raft_binary
ok github.com/hashicorp/vault/vault/external_tests/raft/raft_binary 20.960s