⚠️ Deprecated Content
The content of this page assumes you are using Helm, which is now deprecated. Please use the Installer instead.
Gitpod schedules workspaces as Kubernetes pods. Each workspace pod requests a certain amount of memory which directly affects how many workspaces are scheduled on a single node. If you want to change the default sizing (~ 8GiB per workspace) you should
Merge the following into your
# workspaceSizing configures the resources available to each workspace. These settings directly impact how # desenly we pack workspaces on nodes where workspacesPerNode = memoryAvailable(node) / memoryRequest. # # Beware: if a workspace exceeds its memoryLimit, some of its processes may be terminated (OOM-kill) which # results in a broken user experience. workspaceSizing: requests: cpu: "1m" memory: "2.25Gi" storage: "5Gi" limits: cpu: "5" memory: "12Gi" dynamic: # Gitpod supports dynamic CPU limiting. We express those limits in "buckets of CPU time" (jiffies where 1 jiffie is 1% of a vCPU). # Each bucket has a limit (i.e. max CPU rate in jiffies/sec, 100 jiffies/sec = 1 vCPU). # # For example: # # three minutes of 5 CPUs: 5 [numCPU] * 100 [jiffies/sec] * (3 * 60) [seconds] = 90000 # - budget: 90000 # limit: 500 # # five minutes of 4 CPUs: 4 [numCPU] * 100 [jiffies/sec] * (5 * 60) [seconds] = 120000 # - budget: 120000 # limit: 400 # # remainder of 2 CPUs where a user has to stay below sustained use of 1.8 CPUs for 5 minutes: # # 1.8 [numCPU] * 100 [jiffies/sec] * (5 * 60) [seconds] = 54000 # - budget: 54000 # limit: 200 # # if there are no buckets configured, the dynamic CPU limiting is disabled. cpu: buckets:  samplingPeriod: "10s" controlPeriod: "15m"
helm upgrade --install -f values.custom.yaml gitpod gitpod.io/gitpod --version=0.10.0to apply the changes.