Deploy with Helm

The aperturedata/aperturedb-helm OCI chart deploys distributed ApertureDB into an existing Kubernetes cluster. The default topology has three database pods, three ZooKeeper pods, two database gateways, two HTTP proxies, one checkpointer, a Web UI, and telemetry.

Prerequisites

  • Kubernetes 1.23 or later, kubectl configured for the intended cluster, and Helm 3.8 or later.

  • An object storage bucket and credentials; see Deploying ApertureDB.

  • A storage class that can provision persistent volumes for ZooKeeper and, when telemetry is enabled, a shared 100Gi monitoring volume.

  • Enough cluster capacity for the configured workloads. By default, each database pod and the checkpointer requests 8Gi of memory, and each Lenz pod requests 1Gi, in addition to the other components.

  • LoadBalancer support for separate database TCP and HTTP/HTTPS services, or a plan to expose NodePort or ClusterIP services instead.

Use a separate namespace per deployment: several resource names, including lenz and lenz-external, are fixed within the namespace.

Configure and install

Save the following as my-values.yaml, replacing the placeholders with credentials, a bucket name, and your hostname:

masterKey:
  password: "<strong-database-password>"

storage:
  type: s3
  bucket: "<your-bucket-name>"
  credentials: '{"access_key":"<access-key>","secret_access_key":"<secret-access-key>","region":"us-west-2"}'

service:
  type: LoadBalancer

http:
  host: aperturedb.example.com
  tls: true

telemetry:
  enabled: true
  authPassword: "<separate-monitoring-password>"
  adminPassword: "<strong-grafana-password>"

For Google Cloud Storage, set storage.type: gs and use the service account JSON as storage.credentials. The HTTP service is included; no ingress controller is required.

Select a published package tag and install:

chmod 600 my-values.yaml
helm install aperturedb oci://registry-1.docker.io/aperturedata/aperturedb-helm \
  --version <package-version> \
  --namespace aperturedb --create-namespace \
  -f my-values.yaml --wait --timeout 30m
kubectl -n aperturedb get pods,svc,pvc

Keep this values file for later upgrades. To inspect all settings for the selected package:

helm show values oci://registry-1.docker.io/aperturedata/aperturedb-helm \
  --version <package-version>

Secrets and TLS

For production, reference existing secrets in the deployment namespace instead of placing credentials in the values file. Set the corresponding reference below and omit its inline password or credentials value:

Existing Kubernetes secrets

Helm value

Required secret key

masterKey.existingSecret

password

storage.existingSecret

credentials (the credential JSON)

telemetry.authPasswordSecret

password

telemetry.adminPasswordSecret

password

tls.existingSecret

tls.crt and tls.key in a kubernetes.io/tls secret

For private workload images, set imagePullSecrets to a list such as [{name: my-registry-secret}]. That secret must exist in the same namespace.

TLS selection, in order of precedence, is an existing secret, inline tls.cert and tls.key, cert-manager via tls.certManager, and an automatically generated self-signed certificate. Supply a trusted certificate covering the hostnames used by both web and TCP clients. The self-signed fallback is for evaluation; clients must explicitly trust it or use an evaluation-only certificate verification override.

Connect and verify

Find the database and HTTP/HTTPS endpoints:

kubectl -n aperturedb get service lenz-external http-proxy

Point the DNS record for http.host at the http-proxy address. Native clients use the separate lenz-external address. The TLS certificate must cover all hostnames used by web and native clients.

  • Web UI: https://aperturedb.example.com/.

  • REST API: https://aperturedb.example.com/api/.

  • Native clients: the lenz-external address, port 55555, with TLS enabled. In-cluster clients can use the lenz service on that port.

  • Grafana: https://aperturedb.example.com/grafana/, with telemetry enabled. Log in as admin using the Grafana password configured for the deployment.

Use admin with the master key in the runnable client check at Connect and verify the deployment. Prometheus uses the separate monitoring bearer password to discover and scrape database metrics through Lenz. See Deployment settings and process boundaries for server and gateway settings.

If pods remain pending, inspect kubectl -n aperturedb describe pod <pod-name> and the PVC status for capacity or storage-class problems. For image pull or startup failures, check pod events and kubectl -n aperturedb logs <pod-name> --all-containers --tail=100. An external address that stays pending usually requires checking your cluster’s LoadBalancer implementation. For HTTP routing problems, check DNS, http.host, and the HTTP service logs.

Endpoints by service type

Choose the native and HTTP endpoint independently: service.type controls lenz-external, while http.service.type controls http-proxy.

LoadBalancer endpoints

Read each Service’s external ingress entries:

kubectl -n aperturedb get service lenz-external -o jsonpath='{.status.loadBalancer.ingress}'
kubectl -n aperturedb get service http-proxy -o jsonpath='{.status.loadBalancer.ingress}'

Use the returned IP address or hostname with the Service port. An empty list can mean provisioning is still in progress; non-LoadBalancer Services have no load-balancer ingress entries.

NodePort endpoints

Read the native node port and HTTP/HTTPS node ports:

kubectl -n aperturedb get service lenz-external -o jsonpath='{.spec.ports[?(@.name=="tcp")].nodePort}'
kubectl -n aperturedb get service http-proxy -o jsonpath='{range .spec.ports[*]}{.name}{"="}{.nodePort}{"\n"}{end}'

Use a reachable cluster node address with those ports, including automatically allocated ports. Point http.host at the node or your external load balancer. The certificate must cover the hostname clients use.

ClusterIP endpoints

In-cluster native clients can connect to lenz.aperturedb.svc.cluster.local on service.port (default 55555). The HTTP Service is http-proxy.aperturedb.svc.cluster.local, with Service ports 80 and, when TLS is enabled, 443. Replace aperturedb with your release namespace and cluster.local with your cluster domain if different. These are Service ports, not node ports. Configure http.host and its DNS for the internal HTTP address, and use a certificate covering the hostname each client uses. ClusterIP addresses are reachable within the cluster.

NodePort access

For direct access through cluster nodes, configure distinct, available HTTP and HTTPS NodePorts and allow them through your firewall:

http:
  host: aperturedb.example.com
  service:
    type: NodePort
    httpNodePort: 30080
    httpsNodePort: 30443

Point http.host at a reachable node. With these example ports, the Web UI is at https://aperturedb.example.com:30443/; append api/ or grafana/ for the API or dashboards. HTTP on port 30080 redirects to that HTTPS address. Choose ports from your cluster’s configured NodePort range, normally 30000–32767. To expose native clients through NodePort as well, set service.type to NodePort and inspect lenz-external for its allocated port.

Behind an external load balancer serving ports 80/443, use NodePort and set http.publicPort to 443. Leave the HTTP and HTTPS NodePorts at zero for automatic allocation. Redirects and Grafana URLs use public port 443. Both database and HTTP Services are always created; use ClusterIP for private access.

Workflow routes and request limits

Use http.routes to expose workflow HTTP services under the deployment’s web hostname. Each route specifies a path prefix, an in-cluster service name, and its HTTP port. For example:

http:
  host: aperturedb.example.com
  routes:
    - path: /notebooks
      host: workflow-notebooks-ingresses
      port: 8888

Paths must be unique and omit the trailing slash; /api and /grafana are reserved. Requests retain their path prefix and query string. WebSockets and streamed responses are supported, and route updates take effect automatically.

http.bodyLimit limits each request body to 256 MiB by default. Oversized requests receive 413. http.timeoutSeconds controls upstream inactivity and defaults to 300 seconds; workflow routes and Grafana allow one hour between response reads. Active requests have up to five minutes to finish during shutdown. Long-lived sessions should reconnect after an update.

Valid replacements in the configured TLS Secret are loaded automatically. Existing connections can finish with the previous certificate; invalid replacements leave the last valid certificate active.

Database scratch volume

Set one dbEphemeralVolume object for both database and checkpointer pods:

dbEphemeralVolume:
  size: 100Gi
  storageClass: premium-rwo

The default null uses the container filesystem. When enabled, every pod gets a separate ReadWriteOnce claim with the capacity specified by size. Database pods mount the claim at /aperturedb/db and checkpointer pods at /checkpointer/db. Omit storageClass, or leave it empty, to use global.storageClass, then the cluster default. The class must support dynamic provisioning. Deleting a pod, including scaling it to zero with inactive, deletes its claim; the underlying volume follows the class’s reclaim policy. Distributed database content remains in the object storage bucket. See generic ephemeral volumes.

Settings reference

Add only the values you want to change to my-values.yaml. Helm merges nested values with the chart defaults, so zookeeper: {minSessionTimeout: 30000} keeps the other ZooKeeper settings. helm show values prints the defaults for a package.

In a key such as <workload>.affinity, replace <workload> with a name from the table in Workloads and storage, for example zookeeper.affinity. The http workload is the HTTP proxy. Scheduling and pod annotations under webui apply to the shared Web UI and telemetry pod.

Network

Network settings

Value

Default

Purpose

service.type

LoadBalancer

Type of the lenz-external native database Service: LoadBalancer, NodePort, or ClusterIP.

service.port

55555

Published native database port.

service.nodePort

""

Fixed node port when service.type is NodePort; empty assigns one automatically.

service.annotations, service.loadBalancerClass, service.loadBalancerSourceRanges

{}, "", []

Load balancer annotations, class, and allowed client CIDRs for lenz-external.

http.host

""

Required web hostname used for redirects, Grafana URLs, and generated certificates.

http.extraHosts

[]

Additional web hostnames, also included in generated and cert-manager certificates.

http.tls

true

Serve HTTPS and redirect HTTP to it; false serves HTTP only.

http.publicPort

0

Public web port behind an external load balancer; 0 derives it from the Service.

http.bodyLimit

268435456

Maximum HTTP request body size in bytes.

http.timeoutSeconds

300

Upstream inactivity timeout for API and Web UI requests.

http.routes

[]

Workflow routes; see Workflow routes and request limits.

http.service.type

LoadBalancer

Type of the http-proxy Service.

http.service.httpNodePort, http.service.httpsNodePort

0, 0

Fixed HTTP and HTTPS node ports; see NodePort access.

http.service.annotations, http.service.loadBalancerClass, http.service.loadBalancerSourceRanges

{}, "", []

Load balancer annotations, class, and allowed client CIDRs for http-proxy.

Native client connections always require TLS; the chart has no equivalent of the Compose LNZ_FORCE_SSL setting.

Credentials, TLS, and images

Credential, TLS, and image settings

Value

Default

Purpose

masterKey.password, masterKey.existingSecret

"", ""

Supply one of these for the database admin password.

storage.type

s3

Use s3 or gs for the distributed deployment’s object storage.

storage.bucket

""

Object storage bucket for database content.

storage.credentials, storage.existingSecret

"", ""

Storage credential JSON, inline or from an existing Secret.

storage.extraEnv

{}

Additional non-secret storage variables for database and checkpointer pods.

tls.existingSecret

""

Existing kubernetes.io/tls Secret for web and native TLS.

tls.cert, tls.key

"", ""

Inline PEM certificate chain and private key.

tls.certManager.enabled, tls.certManager.issuerRef.name, tls.certManager.issuerRef.kind

false, "", ClusterIssuer

Issue the certificate with cert-manager.

imagePullSecrets

[]

Existing registry Secrets for private images.

registryCredentials.enabled, registryCredentials.registry, registryCredentials.username, registryCredentials.password

false, https://index.docker.io/v1/, "", ""

Create the aperturedb-registry pull Secret from inline credentials.

image.pullPolicy

IfNotPresent

Pull policy for ApertureData images.

See Secrets and TLS for Secret keys and certificate precedence. The chart version selects matching versions of all ApertureData service images.

Capacity and logging

The package sets a 4096 MiB native message limit and an 8192 MiB shared retained-byte budget (ADB_QUERY_QUEUE_BYTES_MB). Concurrent requests and responses share that budget. Use the database extra-environment settings to adjust both limits together; see Configuration File for required headroom.

Capacity and logging settings

Value

Default

Purpose

aperturedb.maxConcurrency

512

Query workers per database process (ADB_QUERY_THREADS) and the fallback gateway concurrency; independent of accepted connections.

lenz.adbMaxConcurrency

null

Independent gateway concurrency override. Omitted or null follows aperturedb.maxConcurrency. An override must be a positive integer.

aperturedb.taskExecutorThreads

100

Database executor threads.

aperturedb.storageMaxGi

4096

Graph-store storage limit in GiB. The chart requires a positive integer.

aperturedb.storageOperationsLimit

100

Concurrent object storage operations per database process.

aperturedb.descriptorsetFlusherInterval

1800

Seconds between descriptor set cleanup operations.

aperturedb.sessionTokenMaxAgeSec

36000

Session token lifetime in seconds.

aperturedb.refreshTokenMaxAgeSec

360000

Refresh token lifetime in seconds.

aperturedb.healthCheck

true

Enable the database health port and use it for readiness checks.

lenz.timeoutFactor

4

Scales the gateway’s network timeouts; raise it for slow clients or very large requests.

<workload>.extraEnv

{}

Additional environment variables for aperturedb, checkpointer, or lenz, such as ADB_HTTPS_* settings.

lenz.workflowsSecret

""

Existing Secret mounted into each gateway at /etc/lenz/workflows.

zookeeper.tickTime

8000

ZooKeeper tick, in milliseconds, from 1 to 10000.

zookeeper.minSessionTimeout

20000

Minimum timeout ZooKeeper can negotiate with a client, in milliseconds.

zookeeper.publishNotReadyAddresses

true

Publish each ZooKeeper pod’s DNS record before readiness so peers can discover each other during startup. Set to false to publish only ready pods.

zookeeper.maxTimeToWaitForEpoch

30000

Milliseconds a leader waits for a quorum to agree on an epoch before a voter returning to election can trigger a new election. Use 0 to allow a retry as soon as a voter returns to election, or -1 to disable early retries.

For example, configure the gateway independently of the database query threads:

aperturedb:
  maxConcurrency: 256
lenz:
  adbMaxConcurrency: 128

This sets LNZ_ADB_MAX_CONCURRENCY to 128 while the database’s ADB_QUERY_THREADS remains 256. Leaving lenz.adbMaxConcurrency unset or null preserves the existing fallback; both values retain their defaults when neither setting is changed.

The chart sets the other database settings listed for Compose, including ADB_LOG_LEVEL=0 and ADB_KVGD_DB_SYNC=false. See Configuration File for database configuration details.

Workloads and storage

Default replica counts and main container resources

Workload

replicas

resources.requests

resources.limits

aperturedb

3

{cpu: "10m", memory: "8Gi"}

{cpu: "64", memory: "128Gi"}

checkpointer

1

{cpu: "10m", memory: "8Gi"}

{cpu: "64", memory: "128Gi"}

http

2

{cpu: "10m", memory: "64Mi"}

{cpu: "1", memory: "512Mi"}

lenz

2

{cpu: "10m", memory: "1Gi"}

{cpu: "64", memory: "32Gi"}

tools

0

{cpu: "25m", memory: "64Mi"}

{cpu: "1", memory: "256Mi"}

webui

1

{cpu: "25m", memory: "32Mi"}

{cpu: "250m", memory: "128Mi"}

zookeeper

3

{cpu: "250m", memory: "1Gi"}

{cpu: "1", memory: "2Gi"}

When telemetry is enabled, each ApertureDB pod runs one Promtail sidecar. Grafana, Loki, and Prometheus run in the shared web pod. Each container has independent CPU and memory settings.

Telemetry container defaults

Helm value

Requests (CPU, memory)

Limits (CPU, memory)

telemetry.promtail.resources

25m, 64Mi

500m, 256Mi

telemetry.grafana.resources

100m, 256Mi

1, 1Gi

telemetry.loki.resources

100m, 256Mi

2, 2Gi

telemetry.prometheus.resources

100m, 256Mi

2, 2Gi

Web UI resources apply to the webui container in the shared web pod. Use an odd replica count for ZooKeeper. Web UI and Tools each accept 0 or 1. Setting webui.replicas to 0 disables the Web UI and its HTTP route. Telemetry can still run in the shared web pod. With tools.replicas at 0, the Tools Deployment stays scaled down. Set it to 1 to run an admin pod with the deployment’s ZooKeeper server list and prefix.

These Helm values override CPU and memory for each ZooKeeper server and allow its pods to run on nodes tainted dedicated=isvi-cpu:NoSchedule. You can set resources and tolerations independently.

zookeeper:
  resources:
    requests:
      cpu: "250m"
      memory: "512Mi"
    limits:
      cpu: "1"
      memory: "2Gi"
  tolerations:
    - key: dedicated
      operator: Equal
      value: isvi-cpu
      effect: NoSchedule
Workload and storage settings

Value

Default

Purpose

inactive

false

Scale all workloads to zero. Managed persistent claims remain; pod-owned generic ephemeral claims are deleted with their pods.

global.storageClass

""

Storage class for the shared 100Gi monitoring claim and fallback for ZooKeeper and database scratch claims. An empty string uses the cluster default.

dbEphemeralVolume

null

Per-pod database/checkpointer claim with required size and optional storageClass; see Database scratch volume for its lifecycle.

zookeeper.persistence.size, zookeeper.persistence.storageClass

1Gi, ""

Requested capacity and storage class for each ZooKeeper pod’s persistent claim.

zookeeper.prefix

db_root

ZooKeeper path for this deployment’s coordination data.

<workload>.podAnnotations

{}

Map of string annotation keys to string values on the pod template.

<workload>.affinity

{}

Kubernetes affinity object. Combine nodeAffinity, podAffinity, and podAntiAffinity as needed.

<workload>.tolerations

[]

List of toleration objects. The fields key, operator, value, and effect take strings. Use an integer tolerationSeconds for a timed NoExecute toleration.

<workload>.nodeSelector

{}

Map of node label keys to the string values required by the workload’s pods.

Monitoring

Monitoring settings

Value

Default

Purpose

telemetry.enabled

true

Deploy Prometheus, Grafana, Loki, and Promtail log collection.

telemetry.retentionDays

28

Metrics and log retention in days.

telemetry.authPassword, telemetry.authPasswordSecret

"", ""

Supply one when telemetry is enabled. The monitoring password authenticates metrics collection.

telemetry.adminPassword, telemetry.adminPasswordSecret

"", ""

Grafana admin password or the name of an existing Kubernetes Secret with key password. When both are supplied, Helm uses the existing Secret.

Grafana, Loki, and Prometheus share the web Pod, alongside the Web UI when it is enabled. Promtail runs in each database Pod and forwards its logs to Loki.

Upgrade and remove

When upgrading older values files, replace webui.enabled with webui.replicas (0 for false, 1 for true). Remove tools.enabled and set tools.replicas to the desired number of running pods, either 0 or 1.

In older values files, move telemetry.grafana.adminPassword to telemetry.adminPassword and telemetry.auth.password to telemetry.authPassword. For existing Secrets, move telemetry.grafana.existingSecret to telemetry.adminPasswordSecret and telemetry.auth.existingSecret to telemetry.authPasswordSecret. Each reference takes precedence over its inline password.

Move a monitoring storage class configured with telemetry.persistence.storageClass to global.storageClass. For an existing monitoring claim, keep its storage class when upgrading.

Review the compatibility notes in Release Notes for your target version, then complete Changing a deployment. The following command applies the chart with your values file:

helm upgrade aperturedb oci://registry-1.docker.io/aperturedata/aperturedb-helm \
  --version <package-version> --namespace aperturedb \
  -f my-values.yaml --wait --timeout 30m

To remove the deployment:

helm uninstall aperturedb --namespace aperturedb

Uninstall removes the active monitoring PVC and may delete its underlying metrics and log data according to the storage reclaim policy. ZooKeeper PVCs created by StatefulSet claim templates remain. The chart does not delete your object storage bucket. Preserve the bucket and retained claims until you have determined which data must be kept for recovery.