big-data/hadoop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
625039ef20
hadoop/hadoop-tools/hadoop-resourceestimator
History
Subru Krishnan 625039ef20 HADOOP-14840. Tool to estimate resource requirements of an application pipeline based on prior executions. (Rui Li via Subru).
2017-10-25 15:51:27 -07:00
..
src HADOOP-14840. Tool to estimate resource requirements of an application pipeline based on prior executions. (Rui Li via Subru). 2017-10-25 15:51:27 -07:00
pom.xml HADOOP-14840. Tool to estimate resource requirements of an application pipeline based on prior executions. (Rui Li via Subru). 2017-10-25 15:51:27 -07:00
README.md HADOOP-14840. Tool to estimate resource requirements of an application pipeline based on prior executions. (Rui Li via Subru). 2017-10-25 15:51:27 -07:00

README.md

Resource Estimator Service

Resource Estimator Service can parse the history logs of production jobs, extract their resource consumption skylines in the past runs and predict their resource requirements for the new run.

Current Status

  • Support Hadoop YARN ResourceManager logs.
  • In-memory store for parsed history resource skyline and estimation.
  • A Linear Programming based estimator.
  • Provides REST interface to parse logs, query history store and estimations.

Upcoming features

  • UI to query history and edit and save estimations.
  • Persisent store implementation for store (either DB based or distributed key-value like HBase).
  • Integrate directly with the Hadoop YARN Reservation System to make a recurring reservation based on the estimated resources.

Refer to the design document for more details.