python - Amazon DynamoDDB: Are there limits on programmatic changes to a table provisioned throughput? -

python - Amazon DynamoDDB: Are there limits on programmatic changes to a table provisioned throughput? -

i processing big set of files follows: first download of them ec2 instance , run script sequentially processes each file , indexes dynamoddb table (each item in table corresponds single line in file). download next batch of files etc.

right adjust provisioned throughput via aws gui console , ends beingness same during both download phase , indexing phase. clearly, suboptimal since while downloading not talk database @ required write throughput during phase 0.

so want programmatically adjust provisioned throughput way when start indexing , way downwards when stop indexing , start downloading.

are there limits on how much can increment write t/p in single request? instance, can alter 5 120? if not, how calculate number of requests , time required adjust t/p value x value y (where x << y)? have similar calculation decreasing t/p?

i using python boto.

thanks

when scaling can double write provision, if @ 5, can 10, wait 20-30 seconds kick in, double 1 time again 20 , on. can scale downwards low like.

as per aws faq:

q: there limit on how much can alter provisioned throughput single request? yes. amazon dynamodb allows alter provisioned throughput level 100% single updatetable api call. if wish increment throughput more 100%, can phone call updatetable again.

for example, if table has 1,000 units of write capacity provisioned, not update table 3,000 single api phone call more maximum allowed alter single updatetable operation. increment throughput 1,000 3,000 units of write capacity, phone call updatetable first double throughput 2,000, phone call updatetable sec time reach 3,000 writes/second.

python amazon-dynamodb throughput