prateekm opened a new pull request #1241: SAMZA-2424: AM should cache and serve serialized Job Model to containers URL: https://github.com/apache/samza/pull/1241 Symptom: AM runs out of memory or gets stuck in GC loop during job deployment. Cause: During startup, containers fetch the JobModel and configs from the job coordinator http endpoint. For jobs with a large number of containers this can mean many concurrent requests to the JC Servlet for the large job model. The servlet currently serializes the JobModel to JSON for every incoming request. This can lead to excessive memory utilization and GC, leading to slow response times, AM running out of memory, or AM getting stuck in GC loop. Changes: JobModelManager now serializes the JobModel when creating the JobServlet. JobServlet now serves this serialized JobModel instead of serializing every time. Tests: Local performance test results using ab and monitoring memory utilization using YourKit: https://docs.google.com/document/d/1Qy4ScC9aLNZaJQuUBvtyoXdilUhkVtYjb5Mzwzv-DV8/edit?usp=sharing API Changes: None Upgrade Instructions: None Usage Instructions: None
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
