using the s3 provider is taking an incredibly long time for simple stack operations such as

pulumi stack ls

,

time

output:

pulumi stack ls  1.57s user 0.41s system 1% cpu 1:39.90 total

there are ~20 stacks in the bucket, which I can't imagine should take this ls command over a minute to find im running pulumi v3.33.2 is there anything i can do to supply more information on this / is this a known issue with an existing workaround?

Is is possible to use AWS S3 Accelerated to speedup execution? Would it be worth it? Not sure if Accelerated also have great impact on latency.

1,212 objects at 30.6 MB under

.pulumi/stacks

in the s3 bucket

time aws s3 ls <s3://bucketname/.pulumi/stacks> --recursive
0.85s user 0.09s system 19% cpu 4.931 total

just under 5 seconds for aws cli to list the files, which i imagine is all

pulumi stack ls

needs, though there does seem to be an awful lot of

.bak

files up there polluting the space too

For

pulumi up

on stable stack, Pulumi backend seems to be twice as fast as S3 backend. 29s vs 59s. Have anybody found a nice way to make pulumi on S3 backend faster?