azure - Performance analysis of U-SQL script -
when run u sql script portal/visual studio follows stages preparing,queued,running,finalizing. happens behind scenes in these stages?will there execution time difference when job run visual studio/portal in dev , production environment? need clock speeds , record time script take in production.ultimately, goal run these scripts data factory activities in production.
i assume there differences since assume dev environment run @ lower resource usage (lower degree of parallelism both between jobs , inside job) production environment. otherwise there should no difference.
note still working on performance if running particular issues, please let know.
the phases following (i missing parts):
preparing: includes compilation, optimization, codegen, preparing execution graph , required resources , putting job queue.
queueing: job sits in queue executed once job @ top of queue , resources available start job. can impacted setting maximal number of jobs can run in parallel (a setting can set "calling" support/us).
running: actual job execution. affected resources: maximal number of parallelism specified on job, network bandwidth, store access (throttling, bandwidth).
finalizing: cleanup , stitching results files, "sealing" table files. can more expensive depending on write data (adl faster wasb example).
Comments
Post a Comment