https://garden.io logo
#🌱|help-and-getting-started
Breaking limits \ thresholds of garden deployments
# 🌱|help-and-getting-started
c

chilly-beard-31233

12/08/2023, 1:44 PM
Hi team, we are trying to find out the breaking limits or threshold for garden deployments so that to figure out whether we can use garden for larger workloads and large integration testing. And also to use garden in larger scale. Can somebody please help me get information regarding this? Do we have any threshold for it? Appreciate your help. Thanks
q

quaint-dress-831

12/08/2023, 1:45 PM
Hi @chilly-beard-31233, are you working at the same company as @salmon-article-93420? Because if so, they asked a similar question yesterday: https://discord.com/channels/817392104711651328/1182237746459328523
c

chilly-beard-31233

12/08/2023, 1:56 PM
Do we have any official document reference? it should have some restriction right? will it consider all the RAM available on machine?
Yeah i do know him, wanted to get more details on this topic.
q

quaint-dress-831

12/08/2023, 2:54 PM
@chilly-beard-31233 users can override the heap size by setting
GARDEN_MAX_OLD_SPACE_SIZE
and
GARDEN_MAX_SEMI_SPACE_SIZE
but that's only using 0.13.21 and above. Capgemini is using a fork of 0.13.13. https://github.com/garden-io/garden/blob/d53cf1d1d5b2cd1e648bd06edccd156048e790fe/garden-sea/src/node.rs#L42
c

chilly-beard-31233

12/08/2023, 2:57 PM
Yeah this was just one sample question i had posted ...but was really looking for knowing the limits or threshold for garden workloads.
and i am not from Capgemini 😉
e.g. limits.memory limits > memory The maximum amount of RAM the workflow pod can use, in megabytes (i.e. 1024 = 1 GB).
q

quaint-dress-831

12/08/2023, 2:59 PM
Consider that
garden
is calling Kubernetes operations on the backend: the boundedness of workloads is on your Kubernetes cluster, as I explained to your coworker.
c

chilly-beard-31233

12/08/2023, 2:59 PM
does this mean POD which needs more memory wont be able to work in such scenarios?
q

quaint-dress-831

12/08/2023, 2:59 PM
That's up to the limits you set on your own pods.
In other words, they're configurable by you, in your Kubernetes manifests.
Does that make sense? Garden isn't setting boundaries or pod resource limits, either your cluster administrator is and/or they're being set in Kubernetes manifests that you may deploy with
garden
.
e.g. from https://thenewstack.io/kubernetes-requests-and-limits-demystified/#:~:text=A%20container%20created%20using%20this%20manifest%20will%20have,long%20as%20it%20doesn%E2%80%99t%20surpass%20the%20128Mi%20limit
Copy code
yaml
apiVersion: v1
kind: Pod
metadata:
   name: example-pod
   namespace: example-namespace
spec:
   containers:
   - name: example-name
     image: repo/example-image
     resources:
       requests:
         memory: "64Mi"
         cpu: "250m"
       limits:
         memory: "128Mi"
         cpu: "500m"
c

chilly-beard-31233

12/08/2023, 3:05 PM
got it.... thank you very much...i think next step for me is to understand the cluster configurations
q

quaint-dress-831

12/08/2023, 3:06 PM
You're welcome. Just a note that if you're using our
container
Deploy action, you can also set these limits inside Garden's configuration files but it's up to you. For more on that route, see https://docs.garden.io/reference/action-types/deploy/container#spec.limits
We recommend not using the
container
Deploy action, as we plan to deprecate it soon. Users are encouraged instead to use our
kubernetes
and
helm
Deploy actions instead.
c

chilly-beard-31233

12/08/2023, 3:08 PM
ohhh...pitty that most of our workloads are now using container deployments.... 😦