/ec2-patch-bench

Benchmark Pulumi PATCH support across a parameter grid

Primary LanguagePythonApache License 2.0Apache-2.0

Measure baseline vs PATCH performance on EC2 across regions.

Usage

pulumi stack new    meta-benchmark
pulumi stack select meta-benchmark
./make-keys.sh
./configure-keys.sh
pulumi up --yes
# grab several coffees
./ssh-grab-results.sh virginia
./ssh-grab-results.sh oregon
./ssh-grab-results.sh mumbai
./ssh-grab-results.sh frankfurt
./ssh-grab-results.sh tokyo
python analyze.py > results.csv

Configuration scrpits

bench.sh           # confgures parameter grid for benchmarks to run
make-bundle.sh     # confgures what code goes into the zip onto EC2 machines
make-keys.sh       # makes a new SSH key pair
configure-keys.sh  # registers keys to `pulumi config`

Running

Regions: oregon, virginia, mumbai, tokyo, frankfurt.

ssh-grab-results.sh <region>          # rsync results EC2 -> ./data
ssh-there           <region>          # open an ssh shell to EC2
scp-there.sh        <region> file.txt # copy file to EC2

Analysis

python analyze.py  # convert logs from data/ into CSV form for analysis