workload_template

# Copyright (c) 2012 YCSB contributors. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License"); you
# may not use this file except in compliance with the License. You
# may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
# implied. See the License for the specific language governing
# permissions and limitations under the License. See accompanying
# LICENSE file.

# Yahoo! Cloud System Benchmark
# Workload Template: Default Values
#
# File contains all properties that can be set to define a
# YCSB session. All properties are set to their default
# value if one exists. If not, the property is commented
# out. When a property has a finite number of settings,
# the default is enabled and the alternates are shown in
# comments below it.
# 
# Use of most explained through comments in Client.java or 
# CoreWorkload.java or on the YCSB wiki page:
# https://github.com/brianfrankcooper/YCSB/wiki/Core-Properties

# Consider one record in a Timeseries Database as following ([] is optional):
# <metric> <timestamp> <value> [<tag1>=<tagvalue1> <tag2>=<tagvalue2> ... <tagN>=<tagvalueN>]
# tagnames will always be TAG1,TAG2,... ,TAGN
# tagvalus will be random strings
# value will be a random double
# metric is defined as 'usermetric'
# timestamp is automatically generated out of long numbers (which represent milliseconds since 1970)

# The name of the workload class to use
workload=com.yahoo.ycsb.workloads.CoreWorkload

# There is no default setting for recordcount but it is
# required to be set.
# The number of records in the table to be inserted in
# the load phase or the number of records already in the
# table before the run phase.
recordcount=1000000

# There is no default setting for operationcount but it is
# required to be set.
# The number of operations to use during the run phase.
operationcount=3000000

# The number of insertions to do, if different from recordcount.
# Used with insertstart to grow an existing table.
#insertcount=

# The begin of the insertion time space (time: Mo 10. Aug 23:10:05 CEST 2015) (inclusive)
insertstart=1439241005000

# The end of the insertion time space (exclusive)
insertend=1439242005000

# Maximum execution time in seconds
#maxexecutiontime=

# How the latency measurements are presented
measurementtype=histogram
#measurementtype=timeseries

# The range of latencies to track in the histogram (milliseconds)
histogram.buckets=1000

# Granularity for time series (in milliseconds)
timeseries.granularity=1000

#### Parameters for CoreWorkload.java ####

# The name of the database metric to run queries against
metric=usermetric

# The number of tags per record
tagcount=3

# The distribution used to choose the length of a tag value
tagvaluelengthdistribution=constant
#tagvaluelengthdistribution=uniform
#tagvaluelengthdistribution=zipfian

# The size of each tag value (in bytes)
tagvaluelength=10

# Prefix for tags (only use A-Z, a-z)
tagprefix=TAG

# Randomly choose tagamount per record between 0 and tagcount if true
randomtagamount=false

# Generate (randomly) predefined set of values for each tag if true
# otherwise if false generate random tagvalues for every record
# (Consider that if you set this to false, than read/sum/avg/scan/count
# will produce strange results, as you then only have 1 record per tag
# and ycsb-ts will do read/sum/scan/count only based on timespan which
# does not reflect the performance of a TSDB in a good way.)
predefinedtagvalues=true

# Amount of randomly generated predefined set of values per Tag
# (Don't set this to high, otherwise you get in the range of 1
# tag per record, see comment on 'predefinedtagvalues' parameter)
predefinedtagvalueamount=100

# Between RUN and LOAD the tags/tagvalues will be temporarly stored in this file
predefinedtagstoragefile=predefinedtags.txt

# Only stores actually used combos, this means that there can't be
# any misses at all while RUN Phase
predefinedtagstoreused=true

# Range between (e.g. sensor-) values can vary
valuemax=10000
valuemin=0

# Filename containing the tag length histogram
# (only used if tagvaluelengthdistribution is "histogram")
#taglengthhistogram=hist.txt

# What proportion of operations are reads
readproportion=1

# What proportion of operations are inserts
insertproportion=0

# What proportion of operations are scans
scanproportion=0

# what proportion of operations should be avg
avgproportion=0

# what proportion of operations should be count
countproportion=0

# what proportion of operations should be sum
sumproportion=0

# which timevalue to use for avg/count/sum,
# zero means return one value over the whole
# timespan used and only use one "bucket" (default: 0)
# other examples:
# 1 ms = return avg/sum/count for every 1 ms in time period used
# 1 s = return avg/sum/count for every 1 s in time perdiod used
timevalue=0

# which timeunit to use for avg/count/sum
# (possible: ms, s, m, h, d)
timeunit=ms

# The distribution of requests across the keyspace (uniform/latest)
requestdistribution=uniform

# On a single scan/avg/count/sum, the maximum period (in milliseconds) to access
maxscanlength=1000

# On a single scan/avg/count/sum, the maximum period (in milliseconds) to access
minscanlength=500

# The distribution used to choose the period (in milliseconds) to access on a scan/avg/count/sum
scanlengthdistribution=uniform

# should there be more than one value per timestamp possible. Has no effect if randomamountperts=tsrandom.
# (e.g two sensors at the same time with different tags?)
maxrecordsperts=10

# tsrandom, random or constant? random: 1 to maxrecordsperts records for every timestamp, constant:
# always maxrecordsperts records per timestamp, tsrandom: randomly choose timestamp and add value
# (when using tsrandom recordamount between 0 to recordcount per timestamp (maxrecordsperts has no effect))
randomamountperts=constant
#randomamountperts=tsrandom
#randomamountperts=random

# when set to true and using randomamountperts = tsrandom, every timestamp will at least have 1 record after Load Phase
# (Benchmarking will be wrong and recordcount must be greater than (insertstart-insertend)
# Consider that it can happen that there are Zero values when using randomamountperts=tsrandom without flooring
randomfloorfilling=false

# for read/scan/avg/count/sum: how many tagvalues should be given for each tag to look for?
# (it will be randomly chosed how many tags are used and for each tag randomly between 1
# and querymaxtagvalues how many tagvalues)
# e.g. (scan TAG1={1,2,3}, TAG2={1.2},... <tag>={<tagvalues>})
querymaxtagvalues=10

# don't use tags for read, as it is quite hopeless (searching for one timestamp
# and matching the right tags -> not really possible)
notagsforread=true

# Read from values from csv file
readfromcsv=false

# Path to csv file
csvfile=values.csv

# Buffer size for csv reader
csvbuffersize=100000

# Description for the workload
description=Workloaddescription

# Size of one timestep in milliseconds
# 1 = 1 ms, 1000 = 1 s,..
timeresolution=1000

##########################################