httpr

README

HTTP Rake - httpr

HTTP Rake - httpr - is a lightweight and flexible HTTP testing tool that is useful for examining and testing HTTP requests without the need for a full-fledged Web server or a proxy. Implemented in Go, httpr is extremely compact, and can be run locally, which is useful in cases when sending HTTP traffic to a third-party hosted solutions is not desireable, or impractical.

It provides several capabilties that are useful when testing distributed HTTP-based interactions:

• Log incoming requests in raw or JSON format
• Simulate latency
• Simulate a transient HTTP failure
• Return a specific HTTP status code to the HTTP client
• Act as a proxy capable of simulating latency and/or transient failures in front of the upstream service

See httpr help for more information

Installing httpr

Running httpr in Docker

To run httpr in Docker:

docker run -p 8081:8081 netbucket/httpr log -e -p

Installing httpr binary

On macOS:

1. Download the macOS binary:

   curl -L0 https://sourceforge.net/projects/httpr/files/release/1.0.0/darwin/httpr -o httpr

2. Make it executalbe:

   chmod 755 httpr

On Linux:

1. Download the Linux binary:

   curl -L0 https://sourceforge.net/projects/httpr/files/release/1.0.0/linux/httpr -o httpr

2. Make it executalbe:

   chmod 755 httpr

On Windows:

1. Download the Windows binary:

   curl -L0 https://sourceforge.net/projects/httpr/files/release/1.0.0/windows/httpr -o httpr.exe

More to come

Work is underway to make httpr also available via brew for macOS, and apt-get for Linux.

Logging HTTP Requests

To log incoming HTTP requests to standard output, use the httpr log command. Note that by default, httpr will start the HTTP server on port 8081. See httpr help log for more options.

Returning a Specific HTTP Status Code

To return a specific HTTP status code back to the HTTP client, use the -r code option. For instance, to return the HTTP Service Unavailable code, use httpr log -r 503

Simulating Latency

To simulate a delay in returning the HTTP response to the client, use the -d millis option. For instance, to simulate 500 millisecond latency, use httpr log -d 500

Simulating Transient HTTP Failures

httpr can make it easy to simulate transient HTTP errors. This is useful when testing HTTP retry logic, or the circuit breaking capabilities in HTTP clients (see https://martinfowler.com/bliki/CircuitBreaker.html). To do that, use the -f option. The -f option supports additional modifiers to specicfy exactly how the transient failures should be simulated. For instance, to simulate a series of 5 transient failures that return HTTP status 503, followed by 10 successful responses with status code 200, use:

httpr log -f --simulate-failure-count=5 --simulate-failure-code=503 --simulate-success-count=10

Note that -f and -d can be used together to simulate latency and transient errors at once.

Proxying to Simulate Latency and Transient Failures

Using httpr, it is easy to simulate latency or transient failures in front of an existing HTTP based endpoint. To do that, use the httpr proxy command. For instance, to log and then proxy HTTP requests to https://www.google.com, while simulating a transient failure, use:

httpr proxy https://www.google.com -f

TLS/HTTPS Support

To start httpr server in HTTPS mode, use the -t option. By default, httpr will generate and use a self-signed certificate, and print the PEM-encoded certificate to the console. To supply your own certificate, use the --tls-cert-file and --tls-key-file options with the -t flag to specify the path/name of the certificate file and the private key file.

To ingore upstream TLS errors when proxying HTTPS requests with httpr proxy, use the -k flag.