lizeyan
commented
2 years ago A，B，C三个系统不管是调用关系还是部署关系都是固定的，所以我们就直接写死了。对于D系统，调用关系也是固定的，我们就分析了一下trace中服务的调用关系，写死在了https://github.com/lizeyan/train-ticket/blob/master/FDG_data_collection/graph.yml 这里；对于部署关系，我们再通过这个脚本去每次动态补充https://github.com/lizeyan/train-ticket/blob/master/FDG_data_collection/run_collect_metrics.py
Hi,
The paper states that "We construct FDGs automatically, which is necessary because FDGs are complex and dynamic (e.g., in microservice systems, pods are dynamically created and deleted), based on the call and deploy- ment component relationships. ". But I notice that FDG has been constructed previously in datasets. Could you please provide some instructions on how to generate FDG by your codes?