devflowinc/embedding-performance

Testing tps for embeddding servers

0

stars

11

commits

Python

primary language

Jan 14, 2024

updated

README

hi

This is meant to test the performance of embedding servers Using terraform only for the purpose of reproducability.

Testing Setup

Create a markdown graph

graph LR
    A[side_car] --> B((splade_server))
    C[side_car] --> D((embedding_server))

Setup terraform

Download terraform cli Download aws cli

run aws configure You will see a prompt similar to this

user@machine $ aws configure
AWS Access Key ID [****************PYVK]: ****PYVK
AWS Secret Access Key [****************duMt]: ****duMt
Default region name [eu-central-1]: 
Default output format [None]: 

user@machine $

To get the ACCESS KEY and SECRET KEY you must create a user, give that user admin perms and request for an access key for cli usage. The begining of this guide here worked for me. https://spacelift.io/blog/terraform-tutorial

Reproduce tests

terraform init
cp <your-ssh-pub-key-location> ./ssh-key.pub
terraform apply

You should see this as your output

side_car_ip = "<your-side_car-ip>"
splade_ip = "<your-splade-server-ip>"

Set up for splade_server


ssh dev@<your-side_car-ip>

apt-get install python3.9 python3-is-python python-venv

git clone https://github.com/arguflow/embedding-performance
cd embedding-performance
python3 -m venv venv
. venv/bin/activate
pip install -r requirements.txt
granian embeddings:app --interface asgi --host 0.0.0.0 --port 7070

The side_car_ip is what we will test from

Set up for side_car

ssh dev@<your-side_car-ip>

sudo apt-get update -y && sudo apt-get install -y luajit
git clone https://github.com/wg/wrk
git clone https://github.com/arguflow/embedding-performance
cd wrk; make
cd ../embedding-performance/

./test.sh <your-splade-server-ip>:7070
./test.sh <your-embedding-server-ip>:7070

To change the machine type

update the server-machine-type in main.tf

To bring down all servers

terraform destroy

Contributors

cdxker

11 commits

devflowinc/embedding-performance

Testing tps for embeddding servers

0

stars

11

commits

Python

primary language

Jan 14, 2024

updated

README

hi

This is meant to test the performance of embedding servers Using terraform only for the purpose of reproducability.

Testing Setup

Create a markdown graph

graph LR
    A[side_car] --> B((splade_server))
    C[side_car] --> D((embedding_server))

Setup terraform

Download terraform cli Download aws cli

run aws configure You will see a prompt similar to this

user@machine $ aws configure
AWS Access Key ID [****************PYVK]: ****PYVK
AWS Secret Access Key [****************duMt]: ****duMt
Default region name [eu-central-1]: 
Default output format [None]: 

user@machine $

To get the ACCESS KEY and SECRET KEY you must create a user, give that user admin perms and request for an access key for cli usage. The begining of this guide here worked for me. https://spacelift.io/blog/terraform-tutorial

Reproduce tests

terraform init
cp <your-ssh-pub-key-location> ./ssh-key.pub
terraform apply

You should see this as your output

side_car_ip = "<your-side_car-ip>"
splade_ip = "<your-splade-server-ip>"

Set up for splade_server


ssh dev@<your-side_car-ip>

apt-get install python3.9 python3-is-python python-venv

git clone https://github.com/arguflow/embedding-performance
cd embedding-performance
python3 -m venv venv
. venv/bin/activate
pip install -r requirements.txt
granian embeddings:app --interface asgi --host 0.0.0.0 --port 7070

The side_car_ip is what we will test from

Set up for side_car

ssh dev@<your-side_car-ip>

sudo apt-get update -y && sudo apt-get install -y luajit
git clone https://github.com/wg/wrk
git clone https://github.com/arguflow/embedding-performance
cd wrk; make
cd ../embedding-performance/

./test.sh <your-splade-server-ip>:7070
./test.sh <your-embedding-server-ip>:7070

To change the machine type

update the server-machine-type in main.tf

To bring down all servers

terraform destroy

Contributors

cdxker

11 commits

Languages

Python

51.9%

HCL

38.6%

Lua

3.9%

Dockerfile

3.4%

Shell

2.2%