Skip to content

Latest commit

 

History

History
123 lines (90 loc) · 6.06 KB

README.md

File metadata and controls

123 lines (90 loc) · 6.06 KB
Raw

mllp-spammer

Summary

Mllp_spammer.py is a load testing MLLP (Minimal Lower Layer Protocol) tool for sending HL7 messages over the TCP protocol.

At this stage there are 3 separate Apps, with the future roadmap being to make one app with the appropriate flags. During my own usage I found this route quicker and less muddy to use.

App Name Description Notes
Mllp_spammer.py The simplest one Non-persistent connections Prints output Dependencies: pip install APScheduler
Mllp_spammer_axiom.py Added logging to Axiom.co Non-persistent connections Sends logs to Axiom via Axiom SDK (Async HTTP req) Ability to send text file in “once” mode Dependencies: pip install axiom-py python-dotenv APScheduler
Mllp_spammer_axiom_persist.py Uses persistent connections Persistent connections Sends logs to Axiom via Axiom SDK (Async HTTP req) Ability to send text file in “once” mode Has a primitive retry mechanism if disconnected. Retries 3 times, waits 2,3,4 seconds, then stops Dependencies: pip install axiom-py python-dotenv APScheduler

The purpose of this App is to test MLLP and HL7 messages, but it can be used to test any TCP connections. It uses the generic Python Sockets library

Example AWS-Mulesoft Architecture

In this Architecture, we run Mllp_spammer.py inside an EC2 machine in the AWS VPC. Our target MLLP receiver is a Mule App inside Cloudhub 2.0 Private Space (which is another VPC). As this is a Private Space, we have no access via the internet and need to set up routing via AWS Transit Gateway. Find more information on setting up AWS Transit Gateway [here].

In the cases where we use Axiom.co to log messages, both the AWS VPC and Cloudhub 2.0 Private Space need outbound access to the internet.

How to Use

  1. Copy the Python file you want to use to your machine.

    a. If you’re using an Axiom version

     i. Sign up to Axiom hobby tier. You’ll get 2 datasets and at least 100GB to play with
 
 ii. Create API Token >> Settings > API tokens > New API Token
 
 iii. Install the dependencies and create “.env” file with the Axiom API_TOKEN=[my_token_value]

  2. Install Python dependencies (Assumes Python version 3+)

  3. Run via command line python3 mllp_spammer.py [flags]

More details below

Scenario 1 - mllp_spammer.py basic usage

Once mode

python3 mllp_spammer.py -host localhost -p 51961 -sps 1 -mode once

This sends the hardcoded MLLP message to localhost:51961 one time. There are 3 required flags -host (Host) = can be localhost, ip address or uri name -p (Port Number) -sps (Sends Per Second) = 1 to 10 is a good range to use, because the App waits for a reply before sending the next message -mode (spam or once) - this is not required but I like to see it. Spam keeps sending and Once sends once

The way the non-persistent connections work is: Socket opened Message sent Socket closed

Inside the App there is a test HL7 message in the ADT format. I’m not a healthcare expert, so I can’t tell you more than that.

You can also try:

python3 mllp_spammer.py -host tcpbin.com -p 4242 -mode once

Tcpbin.com is a site that reflects back what you sent via tcp on port 4242.

Spam mode

python3 mllp_spammer.py -host localhost -p 51961 -sps 1 -mode spam

Spam mode runs a blocking scheduler from the APScheduler library and runs indefinitely. In order to stop the App when it’s running in the foreground, press Ctrl+C

A blocking scheduler uses 1 thread and waits for a reply before sending the next message. If there’s a slow reply and the scheduler tries to send, you’ll get an error message and it’ll continue to try to send the next message.

There’s currently a 5 second timeout to receive a reply for a message. This is hardcoded, so feel free to change it in the code. Future work is for me to add this as a flag

Running in the background on Linux

nohup [command] &

Runs in the background via nohup, it writes to a nohup.out file, which can grow quickly if left unattended

nohup [command] > file.log &

Runs in the background via nohup, it writes to a file called “file.log”, which can grow quickly if left unattended

nohup [command] > /dev/null 2>&1 &

Runs in the background via nohup and writes no logs

nohup python3 mllp_spammer.py -host localhost -p 51961 -sps 1 -mode spam > file.log &

Runs in mllp_spammer.py in the background and writes to file.log in the same folder

cat /dev/null > nohup.out

Clears the nohup.out file. You can do this while mllp_spammer is still writing to it

Terminating background running processes on Linux

ps -ef | grep python

Views the running processes for python

pkill -9 -f mllp_spammer.py

Closes all running instances of the script, use the script name that is appropriate

kill [id(s)]

Terminate one or more processes, separate ids with spaces

Appendix

To Fix At Some Time:

  • Combine Non-persistent and persistent versions into 1 with flags for options
  • Non-persistent - Hardcoded 5 second timeout for socket connection. This means if no reply occurs in 5 seconds it will disconnect. This should be a flag