SearchVarnish Helm Chart
string
A string. While the YAML spec does not require strings to be quoted, it is highly recommended to quote strings
to prevent YAML type coercion (e.g., values such as country: NO are treated as country: false by YAML).
Example:
key1: "value"
# or multi-line
key2: |
value
number
A number.
key1: 42
# or float
key2: 3.14
boolean
A true or false. While the YAML spec also treats keywords such as “yes” and “no” as true and false, respectively, it is highly recommended to use explicit true and false as the value for maintainability.
key1: true
object
A pair of key value.
key1:
subkey1: "string"
subkey2: 3.14
subkey3: true
# alternatively, using JSON syntax
key2: { "subkey1": "string", "subkey2": 3.14, "subkey3": true }
array of objects
An array of objects.
key1:
- name: "value1"
subkey: "value"
- name: "value2"
subkey: "value"
array of strings
An array of strings.
key1:
- "string1"
- "string2"
- "string3"
# alternatively, using JSON syntax
key1: ["string1", "string2", "string3"]
template string
A pair of key value as a string. Template functions exposed by Helm are available in this type.
key1: |
subkey1: {{ .Release.Name | quote }}
subkey2: "hello, world"
# in array of objects
key2: |
- name: {{ .Release.Name | quote }}
subkey1: "hello, world"
Overrides the name of the chart (without the release name). For example, setting nameOverride to “hello” would produce
a workload named “release-name-hello”. Containers within a pod derive their name from this setting. By default,
the name of the chart is used (i.e., “varnish-enterprise”)
Overrides the full name of the chart (with the release name). This setting allows overriding both release name and
a workload name altogether. For example, setting fullnameOverride to “hello” would produce a workload named
“hello”. By default, a composition of a Helm release name and the name of the chart is used
(i.e., “release-name-varnish-enterprise”).
Applies extra annotations to all workloads. The value can be set as either an object or a template string. Workload annotations can be used to for applying additional metadata or for integrating with external tooling. The annotations specified here will be applied to the workload itself.
To apply annotations on a Pod, use global.podAnnotations. To apply annotations on a specific workload,
use server.annotations.
An array of object that conforms to the Kubernetes imagePullSecrets definition.
When set, each item in an array must consist of an object with a key name referencing the Kubernetes
secret
For example:
global:
imagePullSecrets:
- name: registry-quay-k7c2f4m2d5
Applies extra labels to all workloads. The value can be set as either an object or a template string. Workload labels can be used to for applying additional metadata or for integrating with external tooling. The labels specified here will be applied to the workload itself.
To apply labels on a Pod, use global.podLabels. To apply labels on a specific workload,
use server.labels.
Applies extra annotations to all Pods. The value can be set as either an object or a template string. Pod annotations can be used to for applying additional metadata or for integrating with external tooling. Annotations specified here will be applied to a Pod itself.
To apply labels on a workload, use global.annotations. To apply labels on a specific Pod,
use server.podAnnotations.
Applies extra labels to all Pods. The value can be set as either an object or a template string. Pod labels can be used to for applying additional metadata or for integrating with external tooling. Labels specified here will be applied to a Pod itself.
To apply labels on a workload, use global.labels. To apply labels on a specific Pod,
use server.podLabels.
An object that conforms to the Kubernetes securityContext definition of a Pod
For example:
global:
podSecurityContext:
fsGroup: 999
This securityContext will be set on all Pods within this chart. For setting securityContext on all containers,
see global.securityContext.
An object that conforms to the Kubernetes resources
definition of a Container. This configuration will resources constraint to all containers.
To specify resources on a single container, use server.resources, server.agent.resources, and
server.varnishncsa.resources.
An object that conforms to the Kubernetes securityContext definition of a Container
For example:
global:
securityContext:
runAsUser: 999
runAsNonRoot: true
This securityContext will be set on all containers within this chart. For setting securityContext on the Pod
itself, see global.podSecurityContext.
An object configuring Varnish Controller Agent access to NATS.
An FQDN to the NATS server. The value is ignored if global.natsServer.internal.enabled is set to true. Required when server.agent.enabled is set to true and global.natsServer.internal.enabled is set to false.
trueConfigures Varnish Agent to use the internal NATS server.
Discovers the internal NATS server within the given namespace. By default, the current namespace is used.
Overrides the full name of the internal NATS server. Required if nameOverride or fullnameOverride
is used in the Varnish Controller Helm Chart. By default, “varnish-controller” is used.
cluster.localOverrides the Kubernetes cluster domain. Required if Kubernetes cluster is configured to use different cluster domain than the default. By default, “cluster.local” is used. In most cases, this value should not be changed.
name: varnish-controller-credentials
key: nats-varnish-password
Sets internal NATS password from an external secret. Internal NATS must be configured separately. See Varnish Controller Helm Chart documentation for more information.
For example:
global:
natsServer:
internal:
passwordFrom:
name: secret-name
key: nats-password
Note: if internal Varnish Controller is deployed on another namespace than the one Varnish Enterprise is deployed to, the varnish-controller-credentials secret needs to be manually copied from the Varnish Controller namespace.
trueCreate a Kubernetes service account to use with a workload.
Applies extra labels to the service account. The value can be set as either an object or a template string.
Applies extra annotations to the service account. The value can be set as either an object or a template string.
Overrides the name of the service account. By default, the full name of the chart is used.
1Specifies the number of replicas to deploy Varnish Enterprise server. The value is ignored if server.kind is “DaemonSet”.
Before v1.6.0, when server.autoscaling.enabled is set to true, this value will be ignored and server.autoscaling.minReplicas must be used instead.
From v1.6.0, when server.autoscaling.enabled is set to true, and server.autoscaling.minReplicas is -, the value of server.replicas will be used. server.replicas is ignored if server.autoscaling.minReplicas is set to a number.
DeploymentSpecifies the type of workload to deploy Varnish Enterprise server. The value can be one of
Deployment, DaemonSet, or StatefulSet depending on usage scenarios (see examples).
Applies extra labels to a workload. The value can be set as either an object or a template string.
Labels specified here will be applied to the workload itself. To apply labels on the Pod, use
server.podLabels.
Applies extra annotations to a workload. The value can be set as either an object or a template string.
workload annotations can be used to for applying additional metadata or for integrating with external tooling.
The annotations specified here will be applied to the workload itself. To apply labels on the Pod, use
server.podAnnotations.
Configures deployment strategy to replace existing Pod with a new one.
This configuration is only available when server.kind is set to Deployment. For StatefulSet and
DaemonSet, see server.updateStrategy.
Configures update strategy for updating Pods when a change is made to the manifest. This configuration
is only available when server.kind is set to StatefulSet or DaemonSet. For Deployment, see
server.strategy.
Note: While both StatefulSet and DaemonSet share the same updateStrategy configuration key, its applicable values are different. See updateStrategy on StatefulSet and [updateStrategy on DaemonSet][k8s-daemonset-updatestrategy.
falseWhether to enables hostNetwork for Varnish Enterprise pod. This will make the Varnish Enterprise pod binds to the port of the host instead of using a container network provided by Kubernetes.
Setting this will also set ClusterFirstWithHostNet to allow the Pod to communicate with Kubernetes DNS.
Whether to enable shared PID namespace between all containers in a Pod. This is useful for a scenario where it is necessary to send a signal to a process across a container.
trueConfigures Varnish to listen for HTTP traffic.
6081Configures the TCP port on which Varnish will listen for HTTP traffic. This port is used for Varnish to
bind to within a container. To change the port exposed via service to other applications, see
server.service.http.port.
falseConfigures Varnish to enable In-Process TLS and listens to HTTPS traffic. When enabled,
server.tls.config must also be configured. Certificates may be mounted into the Pod using
server.extraVolumeMounts and server.extraVolumes.
This port is used by HTTP service as a targetPort.
6443Configures the TCP port on which Varnish will listen for HTTPS traffic. This port is used for Varnish to
bind to within a container. To change the port exposed via service to other applications, see
server.service.https.port.
This port is used by HTTPS service as a targetPort.
The TLS configuration. See also In-Process TLS for configuration options. It is recommended
to reference to a port in a configuration using {{ .Values.server.tls.port }} template string for
services integration.
For example:
server:
tls:
config: |
frontend = {
host = ""
port = "{{ .Values.server.tls.port }}"
}
pem-file = "/certs/example.com.pem"
127.0.0.1Configures the address for Varnish management interface.
6082Configures the port for Varnish management interface.
Sets an additional environment variable for Varnish Enterprise container.
Can be set as an array of objects:
extraEnvs:
- name: MY_ENVIRONMENT_VARIABLE
value: my_value
Or an object:
extraEnvs:
MY_ENVIRONMENT_VARIABLE: my_value
Or a template string:
extraEnvs: |
- name: MY_ENVIRONMENT_VARIABLE
value: my_value
An array of extra ports for Varnish to listen to.
For example:
extraListens:
- name: proxy
address: "0.0.0.0"
port: 6888
proto: "PROXY"
- name: proxy-sock
path: "/tmp/varnish-proxy.sock"
user: "www"
group: "www"
mode: "0700"
proto: "PROXY"
The name of the listen. This name will be accessible in VCLs via local.socket.
The protocol of the listen. Must be one of PROXY or HTTP. Default to HTTP if not set.
server.extraListens[].path is setThe port to listens to. Only applicable for TCP listens.
The address to listens to. Only applicable for TCP listens.
server.extraListens[].port is setThe path of UNIX domain socket to listens as. Only applicable for UNIX domain socket.
The user owning the UNIX domain socket. Only applicable for UNIX domain socket.
The group owning the UNIX domain socket. Only applicable for UNIX domain socket.
The file mode octet for the UNIX domain socket. Only applicable for UNIX domain socket.
120Sets the default Time To Live (in seconds) for a cached object.
50Sets the minimum number of worker threads in each pool. See also varnishd documentation.
1000Sets the maximum number of worker threads in each pool. See also varnishd documentation.
120Sets the threshold in seconds where idle threads are destroyed after least this duration.
[]Sets the extra arguments to the Varnish Enterprise.
[]An array of objects that conform to the Kubernetes initContainers definition of a Pod.
This can be used to run initialization tasks before varnishd starts. Note that initContainers
cannot be changed once it is applied. To update this value after the initial deploy, uninstall
Varnish Enterprise Helm Chart from the cluster and reinstall. The value can be set as either an array
of objects or a template string.
[]An array of objects that conforms to the Kubernetes containers definition of a Pod. This can be used to add a sidecar container to varnishd. The value can be set as either an array of objects or a template string.
An array of objects that conforms to the Kubernetes VolumeClaimTemplates
definition of a StatefulSet workload. This configuration is only available when server.kind is
set to StatefulSet.
An array of objects that conforms to the Kubernetes volumeMounts
definition of a Container. This configuration is used to mount extra volumes defined in
server.extraVolumes into the Varnish Enterprise container. The value can be set as either
an array of objects or a template string.
An array of objects that conforms to the Kubernetes volumes definition of a Pod.
This configuration is used to define volumes to be used in server.extraVolumeMounts, or
within server.extraContainers, or within server.extraInitContainers. The value can be set as
either an array of objects or a template string.
server.agent.enabled is true and server.secretFrom isn’t setSets the Varnish secret for accessing the varnishd admin interface. Either this value or
server.secretFrom can be set.
server.agent.enabled is true and server.secret is not setSets the Varnish secret from an external Kubernetes secret for accessing the varnishd admin
interface. Either this value or server.secret can be set.
For example:
server:
secretFrom:
name: secret-name
key: varnish-secret
server.agent.enabled is trueA VCL configuration for Varnish Enterprise.
Before v1.6.0, required if Varnish Controller Agent is not used.
From v1.6.0,
For example:
server:
vclConfig: |
vcl 4.1;
backend default {
.host = "www.example.com";
.port = "80";
}
sub vcl_backend_fetch {
set bereq.http.Host = "www.example.com";
}
Extra VCL configuration where a filename as a key and template string as a value. The path to store
server.vclConfigs will be relative to that of server.vclConfigPath. For example, given the
following configuration:
server:
vclConfigPath: "/etc/varnish/default.vcl"
vclConfigs:
extra.vcl: |
vcl 4.1;
backend default {
.host = "127.0.0.1";
.port = "8090";
}
The file will be saved as /etc/varnish/extra.vcl.
If the filename in server.vclConfigs matches the name in server.vclConfigPath, it will be treated
in the same way as server.vclConfig. In this case, server.vclConfig must not be set.
For example:
server:
vclConfigPath: "/etc/varnish/default.vcl"
# This is effectively the same as setting server.vclConfig: "..."
vclConfigs:
default.vcl: |
vcl 4.1;
backend default {
.host = "127.0.0.1";
.port = "8090";
}
# In this case, vclConfig must be unset.
vclConfig: ""
/etc/varnish/default.vclA path to the main VCL configuration. This configuration affects the location where server.vclConfig
will be saved to, as well as the VARNISH_VCL_CONF environment variable in the Varnish Enterprise
container.
A CLI command file for running management commands when varnishd is launched.
For example, when loading extra VCL in multi-tenancy mode. See also
External VCL ConfigMap.
For example:
server:
cmdfileConfig: |
vcl.load vcl_tenant1 /etc/varnish/tenant1.vcl
vcl.label label_tenant1 vcl_tenant1
vcl.load vcl_main /etc/varnish/main.vcl
vcl.use vcl_main
/etc/varnish/cmds.cliA path to the CLI command file. This configuration affects the location
where server.cmdfileConfig will be saved to, as well as the -I argument in the Varnish
Enterprise container when server.cmdfileConfig is not empty.
Sets a base URL for Varnish Controller and Varnish Controller Router when performing a
HTTP-based routing. If unset, the URL will infer from a Pod IP and server.http.port.
quay.io/varnish-software/varnish-plusSets the repository for Varnish Enterprise image.
IfNotPresentSets the imagePullPolicy for the Varnish Enterprise image. This can be one of Always, Never, or IfNotPreset.
Sets the tag for the Varnish Enterprise image. By default, this is set to the same application
version as in the Varnish Enterprise Helm Chart. If the tag is set to non-exact versions
(such as “latest”, or “6.0”), make sure to set server.image.pullPolicy to “Always” to make
sure the image is always updated.
Applies extra annotations to the Pod. The value can be set as either an object or a template string.
Pod annotations can be used to for applying additional metadata or for integrating with external tooling.
Annotations specified here will be applied to the Pod. To apply labels on a workload, use
server.annotations.
Applies extra labels to the Pod. The value can be set as either an object or a template string.
Labels specified here will be applied to the Pod itself. To apply labels on a workload, use
server.labels.
An object that conforms to the Kubernetes securityContext definition of a Container
For example:
server:
securityContext:
runAsUser: 999
This securityContext will be set on the Varnish Enterprise container. For setting securityContext
on the Pod itself, see global.podSecurityContext. For setting securityContext to all containers,
see global.securityContext.
An object that conforms to the Kubernetes startupProbe definition of a Container
For example:
server:
startupProbe:
initialDelaySeconds: 5
periodSeconds: 10
timeoutSeconds: 1
successThreshold: 1
failureThreshold: 3
Uses Kubernetes httpGet probe instead of TCP probe. Port will be automatically injected. It is possible to provide extra configuration options that conforms to the Kubernetes httpGet probe definition.
For example, to enable httpGet probe mode:
server:
startupProbe:
httpGet:
To specify path and add extra headers:
server:
startupProbe:
httpGet:
path: "/healthz"
httpHeaders:
- name: X-Health-Check
value: k8s
Varnish Helm Chart doesn’t provide a default health check endpoint, so it is necessary to configure Varnish to return 200 OK for this endpoint via a VCL. For example:
server:
vclConfig: |
vcl 4.1
backend default {
.host = "www.example.com";
.port = "80";
}
sub vcl_recv {
if (req.url ~ "^/healthz(/.*)?$") {
# or pass to backend, etc.
return (synth(200, "OK"));
}
}
initialDelaySeconds: 30
periodSeconds: 10
timeoutSeconds: 5
successThreshold: 1
failureThreshold: 3
An object that confirms to the Kubernetes readinessProbe definition of a Container.
Uses Kubernetes httpGet probe instead of TCP probe. Port will be automatically injected. It is possible to provide extra configuration options that conforms to the Kubernetes httpGet probe definition.
See also server.startupProbe.httpGet.
initialDelaySeconds: 30
periodSeconds: 10
timeoutSeconds: 5
successThreshold: 1
failureThreshold: 3
An object that confirms to the Kubernetes readinessProbe definition of a Container.
Uses Kubernetes httpGet probe instead of TCP probe. Port will be automatically injected. It is possible to provide extra configuration options that conforms to the Kubernetes httpGet probe definition.
See also server.startupProbe.httpGet.
An object that conforms to the Kubernetes resources definition of a Container. This configuration can be used to limit resources consumed by the Varnish Enterprise container.
falseWhen server.replicas is set to 1, use Helm Chart release name as Varnish Enterprise server name.
An object that conforms to the Kubernetes nodeSelector definition of a Pod. This configuration is used to select a node to schedule a Pod to. The value can be set as either an object or a template string.
An object that conforms to the Kubernetes tolerations definition of a Pod. This configuration is used to allow the Pod to be scheduled to nodes with specific taints. The value can be set as either an array of strings or a template string.
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchLabels:
app.kubernetes.io/name: {{ include "varnish-enterprise.name" . }}
app.kubernetes.io/instance: {{ .Release.Name }}
topologyKey: kubernetes.io/hostname
This configuration is used to fine-grain control the scheduling of the Pod. By default, this is set to ensure all Varnish Enterprise Pods are always run in a different node. To disable this behavior, set to empty string (""). The value can be set as either an object or a template string.
An object for configuring Massive Storage Engine.
-Enables Massive Storage Engine. Only one of server.mse.enabled or server.mse4.enabled must be enabled.
When set to a string “-”, server.mse.enabled will be set to an opposite value of server.mse4.enabled (i.e. automatically disabling MSE when MSE4 is enabled, and vice versa)
Sets the amount of memory to use for MSE. The value can be set as either a percentage (e.g., “80%”), or bytes (“512m”). If cgroups resource limits are used (i.e., server.resources), the percentage is calculated from the resource limit that is assigned to the container.
An MSE configuration. By default, the value is unset and MSE is run in a memory-only mode. See also the examples in MSE Quick Start for more information on how to configure MSE.
If server.mse.persistence.enabled is true, it is recommended to reference to MSE path using {{ .Values.server.mse.persistence.mountPath }} template string in the configuration.
For example:
server:
mse:
config: |
env: {
id = "env";
memcache_size = "auto";
books = ( {
id = "book1";
directory = "{{ .Values.server.mse.persistence.mountPath }}/book1";
database_size = "1G";
stores = ( {
id = "store";
filename = "{{ .Values.server.mse.persistence.mountPath }}/store1.dat";
size = "9G";
} );
} );
};
falseEnables MSE persistence mode using PersistentVolumes. This configuration is only available when server.kind is set to StatefulSet. See the using StatefulSet documentation for more information.
/var/lib/mseSets the mount path inside the container for PersistentVolume.
Applies extra labels to the PersistentVolume. The value can be set as either an object or a template string.
Applies extra annotations to the PersistentVolume. The value can be set as either an object or a template string.
["ReadWriteOnce"]Sets the AccessMode of the PersistentVolumes. The value is an array of ReadWriteOnce, ReadOnlyMany, or ReadWriteMany. Supports for AccessModes depends on PersistentVolume’s backing storage engine.
Sets the storage class name of a PersistentVolume for choosing which PersistentVolume’s backing storage engine is used. By default, the default PersistentVolume’s backing storage engine is used.
10GiSets the storage volume size to request for. This value should not be less than database size and store size in the MSE configuration.
1%Sets the database size for MSE store. Can be set as a percentage (e.g. 1%) or a literal size (e.g. 10Gi). When the value is set as a percentage, the value will be relative to server.mse.persistence.storageSize.
91%Sets the store size for MSE store. Can be set as a percentage (e.g. 1%) or a literal size (e.g. 10Gi). When the value is set as a percentage, the value will be relative to server.mse.persistence.storageSize.
An object for configuring [Massive Storage Engine 4][mse4].
falseEnables Massive Storage Engine 4. Only one of server.mse.enabled or server.mse4.enabled must be enabled.
Sets the amount of memory to use for MSE4. The value can be set as either a percentage (e.g., “80%”), or bytes (“512m”). If cgroups resource limits are used (i.e., server.resources), the percentage is calculated from the resource limit that is assigned to the container.
An MSE configuration. By default, the value is unset and MSE4 is run in a memory-only mode.
If server.mse4.persistence.enabled is true, it is recommended to reference to MSE path using {{ .Values.server.mse4.persistence.mountPath }} template string in the configuration.
For example:
server:
mse4:
config: |
env: {
books = ( {
id = "book";
filename = "{{ .Values.server.mse4.persistence.mountPath }}/book";
size = "1G";
stores = ( {
id = "store";
filename = "{{ .Values.server.mse4.persistence.mountPath }}/store";
size = "9G";
} );
} );
};
falseEnables MSE4 persistence mode using PersistentVolumes. This configuration is only available when server.kind is set to StatefulSet. See the using StatefulSet documentation for more information.
/var/lib/mse4Sets the mount path inside the container for PersistentVolume.
Applies extra labels to the PersistentVolume. The value can be set as either an object or a template string.
Applies extra annotations to the PersistentVolume. The value can be set as either an object or a template string.
["ReadWriteOnce"]Sets the AccessMode of the PersistentVolumes. The value is an array of ReadWriteOnce, ReadOnlyMany, or ReadWriteMany. Supports for AccessModes depends on PersistentVolume’s backing storage engine.
Sets the storage class name of a PersistentVolume for choosing which PersistentVolume’s backing storage engine is used. By default, the default PersistentVolume’s backing storage engine is used.
10GiSets the storage volume size to request for. This value should not be less than database size and store size in the MSE4 configuration.
1%Sets the database size for MSE4 store. Can be set as a percentage (e.g. 1%) or a literal size (e.g. 10Gi). When the value is set as a percentage, the value will be relative to server.mse4.persistence.storageSize.
91%Sets the store size for MSE4 store. Can be set as a percentage (e.g. 1%) or a literal size (e.g. 10Gi). When the value is set as a percentage, the value will be relative to server.mse4.persistence.storageSize.
An object for configuring HorizontalPodAutoscaling.
falseEnables the HorizontalPodAutoscaling with the Varnish Enterprise Pod. server.replicas
is ignored if autoscaling is enabled.
Configures the behavior attribute of HorizontalPodAutoscaling of the Varnish Enterprise Pod.
1-Sets the minimum number of replicas to always keep running.
From v1.6.0, when set to -, the value will inherit to that of server.replicas.
100Sets the maximum number of replicas to run at most.
Sets the HorizontalPodAutoscaling metrics. The value can be set as either an object or a template string.
An object for configuring PodDisruptionBudget.
falseEnables PodDisruptionBudget.
server.pdb.enabled is true and server.pdb.maxUnavailable is not setSets the number or percentage of pods that must be available after the eviction.
server.pdb.enabled is true and server.pdb.minAvailable is not setSets the number or percentage of pods that can be unavailable after the eviction.
noneSets the method for delaying shutdown of Varnish Enterprise. Can be one of none, sleep, mempool, or shutdown_delay (Varnish Enterprise 6.0.13r2 and newer).
This is useful for rolling updates or when scaling down, where the load balancer component may still attempt to forward requests to Varnish Enterprise instance when Varnish Enterprise has already received a SIGHUP, leading to HTTP 502 errors.
Multiple delayed shutdown methods are available:
sleep will simply sleep in the preStop for server.delayedShutdown.sleep.seconds seconds. This is the simplest way to delay the shutdown of Varnish Enterprise when the time taken for the load balancer component to stop forwarding traffic to Varnish Enterprise is largely static.mempool polls varnishstat for MEMPOOL.sess to go down to zero. This allows for the quickest shutdown but requires tuning server.delayedShutdown.mempool.pollSeconds to the traffic patterns, as MEMPOOL.sess may become 0 momentary resulting in an early shutdown.shutdown_delay similar to sleep, but utilizes Varnish Enterprise’s shutdown_delay feature (available since 6.0.13r2) to perform the wait in-engine instead of relying on preStop and an external command.When server.delayedShutdown.method is set to a value other than “none”, server.terminationGracePeriodSeconds must also be set.
90Sets the number of seconds to sleep before SIGHUP is sent to Varnish Enterprise. Only applicable when server.delayedShutdown.method is set to “sleep”.
1Sets the number of seconds to poll for MEMPOOL.sess stats. Only applicable when server.delayedShutdown.method is set to “mempool”.
This value needs to be tuned to the traffic patterns. If the time between each request took more than server.delayedShutdown.mempool.pollSeconds, then Varnish Enterprise could be signaled to SIGHUP too early, leading to an HTTP 502 error on the load balancer (as Varnish Enterprise already stopped accepting connection).
5Sets the number of seconds to wait after MEMPOOL.sess turned zero. Only applicable when server.delayedShutdown.method is set to “mempool”.
Sets the number of seconds for varnishd to wait before terminating itself. This is similar to sleep, but use an in-engine method of waiting instead of relying on preStop and an external command. Only applicable when server.delayedShutdown.method is set to “shutdown_delay”.
Setting this value is equivalent of setting -p shutdown_delay=<number_of_seconds> -p shutdown_close=off as an argument to varnishd.
Sets the termination grace period seconds to wait for Varnish Enterprise to shut down. Once this number of seconds is reached without Varnish Enterprise Pod shutting down, Kubernetes will send SIGKILL to the Pod. When server.delayedShutdown is used, this value should be set to a higher value than the maximum time (in seconds) it took for Varnish Enterprise to fully shutdown after a delayed shutdown.
The value to set here depends on the cluster setup. For example, an AWS Load Balancer may still forward requests to Varnish for deregistration_delay.timeout_seconds which is set to 300 seconds by default. If server.delayedShutdown took 30 seconds, then server.terminationGracePeriodSeconds should be set to at least 330 seconds in order to have a zero-downtime during a rolling update or scaling down.
An object for configuring varnishncsa.
trueEnables HTTP request logging via varnishncsa.
Sets an additional environment variable for varnishncsa container.
Can be set as an array of objects:
extraEnvs:
- name: MY_ENVIRONMENT_VARIABLE
value: my_value
Or an object:
extraEnvs:
MY_ENVIRONMENT_VARIABLE: my_value
Or a template string:
extraEnvs: |
- name: MY_ENVIRONMENT_VARIABLE
value: my_value
An array of objects that conforms to the Kubernetes volumeMounts
definition of a Container. This configuration is used to mount extra volumes defined in
server.extraVolumes into the varnishncsa container. The value can be set as either an
array of objects or a template string.
-Sets the repository for Varnish Enterprise image for use with varnishncsa. The Varnish image used
here must be the same version as the Varnish Enterprise server. Set this to “-” to inherit the value
of server.image.repository.
-Sets the imagePullPolicy for the Varnish Enterprise image for use with
varnishncsa. This can be one of Always, Never, or IfNotPreset. Set this to “-” to inherit the value
of server.image.pullPolicy.
-Sets the tag for the Varnish Enterprise image for use with varnishncsa. The Varnish image used
here must be the same version as the Varnish Enterprise server. Set this to “-” to inherit the value
of server.image.tag.
An object that conforms to the Kubernetes securityContext definition of a Container.
For example:
server:
securityContext:
runAsUser: 999
This securityContext will be set on the varnishncsa container. For setting securityContext
on the Pod itself, see global.podSecurityContext. For setting securityContext to all containers,
see global.securityContext.
[]Sets the extra arguments to Varnish NCSA.
An object that conforms to the Kubernetes startupProbe definition of a Container.
initialDelaySeconds: 5
periodSeconds: 10
timeoutSeconds: 1
successThreshold: 1
failureThreshold: 3
An object that confirms to the Kubernetes readinessProbe definition of a Container.
initialDelaySeconds: 30
periodSeconds: 10
timeoutSeconds: 5
successThreshold: 1
failureThreshold: 3
An object that confirms to the Kubernetes livenessProbe definition of a Container.
An object that conforms to the Kubernetes resources definition of a Container. This configuration can be used to limit resources consumed by the varnishncsa container.
An object for configuring the Varnish OTEL container for OpenTelemetry observability integration.
falseEnables the Varnish OTEL container for OpenTelemetry integration. When
enabled, the container runs /usr/bin/varnish-otel to export logs and
metrics from Varnish to an OTLP (OpenTelemetry Protocol) endpoint.
-Sets the repository for Varnish Enterprise image for use with Varnish
OTEL. The Varnish image used here must be the same version as the Varnish
Enterprise server. Set this to “-” to inherit the value of
server.image.repository.
-Sets the imagePullPolicy for the Varnish
Enterprise image for use with Varnish OTEL. This can be one of Always,
Never, or IfNotPreset. Set this to “-” to inherit the value of
server.image.pullPolicy.
-Sets the tag for the Varnish Enterprise image for use with Varnish OTEL.
The Varnish image used here must be the same version as the Varnish
Enterprise server. Set this to “-” to inherit the value of
server.image.tag.
An object that conforms to the Kubernetes securityContext definition of a Container.
For example:
server:
otel:
securityContext:
runAsUser: 999
This securityContext will be set on the Varnish OTEL container. For
setting securityContext on the Pod itself, see global.podSecurityContext.
For setting securityContext to all containers, see global.securityContext.
An object that conforms to the Kubernetes resources definition of a Container. This configuration can be used to limit resources consumed by the Varnish OTEL container.
Sets environment variables for the Varnish OTEL container. This is typically used to configure OpenTelemetry exporters and other OTEL settings. Most variables have sensible defaults; see the Varnish OTEL Configuration documentation.
Can be set as an object:
server:
otel:
enabled: true
env:
OTEL_EXPORTER_OTLP_ENDPOINT: "http://otel-collector:4317"
OTEL_SERVICE_NAME: "varnish"
Or as an array of objects:
server:
otel:
env:
- name: OTEL_EXPORTER_OTLP_ENDPOINT
value: "http://otel-collector:4317"
- name: OTEL_SERVICE_NAME
value: "varnish"
Or as a template string:
server:
otel:
env: |
- name: OTEL_EXPORTER_OTLP_ENDPOINT
value: "http://otel-collector:4317"
- name: OTEL_SERVICE_NAME
value: {{ .Release.Name | quote }}
For a complete list of OpenTelemetry environment variables, see the OpenTelemetry Environment Variable Specification.
An object for configuring Varnish Controller Agent.
falseEnables the Varnish Controller Agent for integration with Varnish Controller.
If Varnish Controller is deployed in a different namespace than Varnish Enterprise,
also set global.natsServer.internal.namespace.
If Varnish Controller is deployed outside the Kubernetes cluster, or is using an
external NATS server not provided by Varnish Controller Helm Chart, set
global.natsServer.internal.enabled to false and configures
global.natsServer.externalAddress.
noneEnable automatic de-registering of Varnish Controller Agent during Pod rollout. The value can be one of none (disabled), or vcli (use VCLI to perform de-registration).
quay.io/varnish-software/varnish-controller-cliSets the repository for Varnish Controller VCLI image. The Varnish Controller VCLI image used here must be the same version as the Varnish Controller. Since this is a different image from Varnish Enterprise, using “-” is unsupported.
IfNotPresentSets the imagePullPolicy for the Varnish Controller VCLI image. This can be one of Always, Never, or IfNotPreset.
Sets the tag for the Varnish Controller VCLI image for automatic de-registration. The Varnish Controller VCLI image used here must be the same version as the Varnish Controller. Since this is a different image from Varnish Enterprise, using “-” is unsupported.
An FQDN to the Varnish Controller API-GW server. The value is ignored if server.agent.autoRemove.vcli.internal.enabled is set to true. Required when server.agent.autoRemove.method is set to “vcli” and server.agent.autoRemove.vcli.internal.enabled is set to false.
Sets the username to connect to Varnish Controller. Required when server.agent.autoRemove.method is set to vcli.
Sets the password to connect to Varnish Controller. Required when server.agent.autoRemove.method is set to vcli and server.agent.autoRemove.vcli.passwordFrom is not set.
name: varnish-controller-credentials
key: varnish-admin-password
Sets the Varnish Controller password from an external secret. Required when server.agent.autoRemove.method is set to vcli and server.agent.autoRemove.vcli.password is not set.
Note: if internal Varnish Controller is deployed on another namespace than the one Varnish Enterprise is deployed to, the varnish-controller-credentials secret needs to be manually copied from the Varnish Controller namespace.
falseConfigures Varnish Controller CLI container to skip verifying the TLS certificate when communicating via HTTPS. This should be set to true if Varish Controller is running with a self-signed certificate.
trueConfigures whether to use the Varnish Controller instance deployed by the Varnish Controller Helm Chart. When server.agent.autoRemove.vcli.internal.enabled is set to true, most configurations should be auto-discovered.
Sets the namespace of the Varnish Controller when Varnish Controller is deployed in a different namespace than that of Varnish Enterprise. When this value is empty, the Varnish Enterprise Helm Chart will use the current namespace for auto-discovery.
Sets the release name of the Varnish Controller when Varnish Controller is deployed with a different release name than “varnish-controller”. When this value is empty, the Varnish Enterprise Helm Chart will use “varnish-controller”.
8080Sets the port for API-GW.
Sets whether to use HTTPS to connect to Varnish Controller API-GW.
An object that conforms to the Kubernetes resources definition of a Container. This configuration can be used to limit resources consumed by the Varnish Controller VCLI container.
Sets an additional environment variable for Varnish Controller VCLI container.
Can be set as an array of objects:
extraEnvs:
- name: MY_ENVIRONMENT_VARIABLE
value: my_value
Or an object:
extraEnvs:
MY_ENVIRONMENT_VARIABLE: my_value
Or a template string:
extraEnvs: |
- name: MY_ENVIRONMENT_VARIABLE
value: my_value
An array of objects that conforms to the Kubernetes volumeMounts definition of a Container. This configuration is used to mount extra volumes defined in server.extraVolumes into the Varnish Controller Agent container. The value can be set as either an array of objects or a template string.
[]Sets the extra arguments to the Varnish Controller Agent.
Sets an additional environment variable for Varnish Controller Agent container.
Can be set as an array of objects:
extraEnvs:
- name: MY_ENVIRONMENT_VARIABLE
value: my_value
Or an object:
extraEnvs:
MY_ENVIRONMENT_VARIABLE: my_value
Or a template string:
extraEnvs: |
- name: MY_ENVIRONMENT_VARIABLE
value: my_value
An array of objects that conforms to the Kubernetes volumeMounts
definition of a Container. This configuration is used to mount extra volumes defined in
server.extraVolumes into the Varnish Controller Agent container. The value can be set
as either an array of objects or a template string.
quay.io/varnish-software/varnish-controller-agentSets the repository for Varnish Controller Agent image. The Varnish Controller image used here must be the same version as the Varnish Controller. Since this is a different image from Varnish Enterprise, using “-” is unsupported.
IfNotPresentSets the imagePullPolicy for the Varnish Controller image. This can be one of Always, Never, or IfNotPreset.
Sets the tag for the Varnish Controller image for use with varnishncsa. The Varnish Controller image used here must be the same version as the Varnish Controller. Since this is a different image from Varnish Enterprise, using “-” is unsupported.
Sets a Varnish Controller private token.
An object that conforms to the Kubernetes securityContext definition of a Container.
For example:
server:
securityContext:
runAsUser: 999
This securityContext will be set on the Varnish Controller container. For setting
securityContext on the Pod itself, see global.podSecurityContext. For setting
securityContext to all containers, see global.securityContext.
debugSets the log level for the Varnish Controller Agent. The value can be one of
debug, info, warning, error, or quiet.
Sets the tags for the Varnish Controller Agent.
Sets the latitude and longitude for Varnish Controller Agent. These values are used to display the location of Varnish instance on a map on the Varnish Controller UI, and for GeoIP routing in Varnish Traffic Router.
Sets the longitute where the server is located.
Sets the latitude where the server is located.
An object that conforms to the Kubernetes resources definition of a Container. This configuration can be used to limit resources consumed by the Varnish Controller Agent container.
An object for configuring Varnish Controller Agent’s state persistency.
falseEnables Varnish Controller Agent state persistency using PersistentVolumes.
This configuration is only available when server.kind is set to StatefulSet. See the
Integration with Varnish Controller Agent documentation
for more information.
Enables Varnish Controller Agent state persistency using a custom volume mounts.
This configuration overrides server.agent.persistence.enabled. See the
Integration with Varnish Controller Agent documentation
for more information.
Applies extra labels to the PersistentVolume. The value can be set as either an object or a template string.
Applies extra annotations to the PersistentVolume. The value can be set as either an object or a template string.
["ReadWriteOnce"]Sets the AccessMode of the PersistentVolumes. The value is an array of ReadWriteOnce, ReadOnlyMany, or ReadWriteMany. Supports for AccessModes depends on PersistentVolume’s backing storage engine.
Sets the storage class name of a PersistentVolume for choosing which PersistentVolume’s backing storage engine is used. By default, the default PersistentVolume’s backing storage engine is used.
512MiSets the storage volume size to request for.
trueConfigures whether to use init container to initialize Varnish Controller Agent state directory symbolic link. This is to allow Varnish Enterprise Pod to have a static path for accessing Varnish Controller Agent state files. When server.initAgent.enabled is false, the Varnish Controller Agent state directory will be discovered using a dynamic environment variable based on Pod name.
Before v1.6.0, the state directory symbolic link is created at /etc/varnish/agent.
From v1.6.0, the state directory symbolic link is created at /etc/varnish/shared/agent.
From v1.8.0, the default will be changed to false.
Sets an additional environment variable for Varnish Controller Agent’s init container.
Can be set as an array of objects:
extraEnvs:
- name: MY_ENVIRONMENT_VARIABLE
value: my_value
Or an object:
extraEnvs:
MY_ENVIRONMENT_VARIABLE: my_value
Or a template string:
extraEnvs: |
- name: MY_ENVIRONMENT_VARIABLE
value: my_value
An object that conforms to the Kubernetes securityContext definition of a Container
For example:
server:
initAgent:
securityContext:
runAsUser: 999
This securityContext will be set on the Varnish Controller Agent’s init container.
For setting securityContext on the Pod itself, see global.podSecurityContext.
For setting securityContext to all containers, see global.securityContext.
An object for configuring Service.
trueEnables the Service for Varnish Enterprise.
Applies extra labels to the Service. The value can be set as either an object or a template string.
Applies extra annotations to the Service. The value can be set as either an object or a template string.
NodePortSets the type of the Service. Can be either CluterIP, LoadBalancer, or NodePort.
See also Accessing Varnish Enterprise in Kubernetes Cluster
Sets a custom Service ClusterIP. This value can be set as either an IP address,
or a literal string “None”. Only applicable when server.service.type is set to ClusterIP. When set to
“None”, Kubernetes will create a Headless Service, skipping the Kubernetes proxying mechanism.
In such case, server.service.http.port and server.service.https.port will be ignored.
Instead, server.http.port and server.tls.port must be used instead.
See also Accessing Varnish Enterprise in Kubernetes Cluster.
Sets a pre-allocated IP address for use in LoadBalancer.
trueEnables HTTP service.
80Sets the port to expose HTTP service.
Sets the port to expose HTTP service directly on the node itself. Only applicable when server.service.type
is set to NodePort. This value must be within the Kubernetes service-node-port-range (default: 30000-32767).
See also Accessing Varnish Enterprise in Kubernetes Cluster.
falseEnables HTTPS service. Requires server.tls.enabled to be enabled.
443Sets the port to expose HTTPS service.
Sets the port to expose HTTPS service directly on the node itself. Only applicable when server.service.type
is set to NodePort. This value must be within the Kubernetes service-node-port-range (default: 30000-32767).
See also [accessing Varnish via NodePort][access-nodeport].
An array of extra services to expose to as a Service.
For example:
extraServices:
- name: "varnish-proxy"
targetPort: 6888
port: 88
Sets the name of the Service.
Sets the target ports that are exposed via server.extraListens.
Sets the port to expose this extra service.
Sets the port to expose this extra service on the node itself. Only applicable when server.service.type
is set to NodePort. This value must be within the Kubernetes service-node-port-range (default: 30000-32767).
An object for configuring Ingress.
falseEnables the Ingress for Varnish Enterprise.
Applies extra labels to the Ingress. The value can be set as either an object or a template string.
Applies extra annotations to the Ingress. The value can be set as either an object or a template string.
Sets the Ingress Class for selecting Ingress controller to use.
PrefixSets the Ingress Path Type for the Varnish Enterprise endpoint. The value can be
either Prefix, Exact, or ImplementationSpecific. The value to use here depends on the Ingress
controller.
Sets the hostname for the Ingress. This hostname is used for routing traffic.
An array of objects that conforms to Ingress TLS.
An array of objects to attach Kubernetes manifests to a workload.
For example:
extraManifests:
- name: clusterrole
data: |
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
name: {{ .Release.Name }}-clusterrole
# ...
The name of the manifest. Only used if extraManifests[].checksum is true.
Whether to attach the manifest’s checksum to that of the workload in order to force an automatic rollout when the manifest is updated.
The full content of the manifest.
Enables clustering and allows pods to shard requests between each others by:
server.service.clusterIP != NoneConfigures the headless service used by pods to discover each others. If empty,
the name defaults to $VARNISH_SERVICE-peers and the service is automatically
created.
Configures the token used by nodes to authorize peer-to-perr communication.
If empty, a $VARNISH_SERVICE-cluster-secret secret is automatically created.
Activates traces.