https://pulumi.com logo
#general
Title
# general
c

crooked-helicopter-55521

03/31/2020, 7:15 PM
Dumb question, can you configure the 10 minute timeout on
pulumi up
? Digging through the docs I can't find anything, which seems hard to belive. Am I missing something? It's been causing problems for a long running change I'm trying to do
c

colossal-room-15708

03/31/2020, 8:38 PM
c

crooked-helicopter-55521

03/31/2020, 8:46 PM
Thanks for the link David. It's close to what I want but a bit of a bummer it has to be configured per resource. I guess what I want doesn't exist... 99% of my operations are quick, but this one operation can take 20 minutes, so I what I want to be able to do is
pulumi up --customTimeout=30m
instead of globally upping the timeouts. Thanks again though, good to know what's possible
w

white-balloon-205

03/31/2020, 9:55 PM
What resources are you using? The underlying providers associated with each resource define their own default timeouts - there is no Pulumi wide timeout at all. So it is necessary to change defaults on a per-resource basis. In general, these timeouts should be designed to be appropriate for the 90% case - so timeouts should not frequently need to be modified.
c

crooked-helicopter-55521

03/31/2020, 9:58 PM
gcp.container.Cluster
is the one specifically in question. What's causing the problem is when I go to upgrade the kubernetes node version, it takes 3-5 minutes a node, and does it sequentially. So, it takes 100x longer than any other normal change. Using the docs David showed me, I think I've got my timeout set high enough to allow that to all complete. My current gameplan is to just remove the custom timeout, and put it back when I next need to upgrade.
👍 1