If running in Hetzner and using Hetzner's cloud load balancers, node
ports need to be allocated for the load balancer to function correctly.
Therefore if RobotLB is enabled, we probably need to assign node ports.
Release note:
[platform] Autodetect if node ports should be assigned to load balancer
services.
Signed-off-by: Timofei Larkin <lllamnyp@gmail.com>
This patch passes Java heap parameters to Keycloak to prevent OOM errors
when the JVM lacks compatibility with cgroups v2 and fails to recognize
container memory requests and limits. A new function is introduced in
cozy-lib to calculate the heap parameters from requests and limits,
setting Xmx to 75% of the memory limit and Xms to the lesser of the
memory request or 25% of the memory limits.
Change log:
[keycloak] Calculate and pass Java heap parameters explicitly to prevent
OOM errors.
[cozy-lib] Introduce helper function to calculate Java heap params based
on memory requests and limits.
Signed-off-by: Timofei Larkin <lllamnyp@gmail.com>
This patch removes the loophole to leave resource requests and limits
unspecified in managed apps. Any of cpu, memory, and ephemeral storage
are now filled in from the resource preset (default or user-specified)
if not explicitly specified in .Values.resources. "none" is no longer an
accepted value in resourcePresets and the primary resources now always
have some explicit value for proper billing and isolation.
Signed-off-by: Timofei Larkin <lllamnyp@gmail.com>
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
* **Refactor**
* Updated the structure of resource presets for improved clarity and
processing.
* Adjusted template logic to streamline resource handling and removed
previous resource limit calculations.
* Modified template parameters to enhance flexibility in resource
processing.
* **Chores**
* Improved internal template invocation for better compatibility with
resource data.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
With this change a request for a virtual machine with 3 vCPUs will
reserve exactly the same amount of physical compute, as a request for a
Clickhouse instance with `{"resources": {"cpu": "3"}}` in its values,
with the scaling factor being KubeVirt's CPU allocation ratio.
Signed-off-by: Timofei Larkin <lllamnyp@gmail.com>
This patch introduces reusable library charts that provide
backward-compatibility for users that specify their resources as
explicit requests and limits for cpu, however this input is processed so
that limits are set equal to requests except for CPU which only gets
requests. Users can now embrace the new form by directly specifying
resources in the first level of nesting (e.g. resources.cpu=100m instead
of .resources.requests.cpu=100m). The order of precedence is top-level,
then requests, then limits, ensuring that nothing will break in terms of
scheduling, however workloads that specified limits much higher than
requests might get a performance hit, now that they cannot use all this
excess capacity. This should only affect memory-hungry workloads in
low-contention environments.
Signed-off-by: Timofei Larkin <lllamnyp@gmail.com>
