-
Install PostgreSQL 14 or later along with psql
https://www.postgresql.org/ -
Clone the repository from Github or Zenodo
https://github.com/TheMonocledHamster/SynConfLoad/
https://zenodo.org/badge/latestdoi/568378464 -
Download the Azure Traces dataset from https://github.com/Azure/AzurePublicDataset/blob/ef8b2517b27357df0b418b6e6ca4efcdeb5117b0/AzureFunctionsDataset2019.md
-
From the dataset, copy 2 files, function_durations_percentiles.anon.d01.csv and invocations_per_function_md.anon.d01 into the source/traces/AzureFunctions/ directory. Your directory structure should look something like this:
-
Ensure that the PostgreSQL Database is properly installed and in operation.
-
Run sudo -u postgres bash source/db/run.sh from the project root directory.
-
The script may take anywhere from 1-3 hours to execute, with a max disk space requirement of around 10 GB.
-
Upon completion, do the following:
-
Install psycopg2 and numpy using
pip install psycopg2-binary
pip install numpy -
Navigate to source/arrival_rates/gen_arrivals.py
-
Change the user to your username
-
Run python gen_arrivals.py
The final data should be populated into sub-folders within the arrival_rates directory, as .csv and .npy files for each SLO bin.