Helm Chart installation
Oracle Backend for Microservices and AI (OBaaS) Helm Charts
This document describes how to deploy OBaaS to an existing Kubernetes cluster using Helm charts.
- Quick Start Decision Guide — Choose the right example values file for your deployment
- Architecture — Two-chart design, components, and cluster layout
- Dependent Helm Chart References — Version-pinned README and values references for bundled dependency charts
- Prerequisites — Required tools and cluster access
- Names and Namespace Conventions — Replace example release names and namespaces with your own values
- Installation Guide — Step-by-step deployment process
- Example Configurations — Common deployment scenarios:
- Default — Quick start with no overrides
- SIDB-FREE Database — In-cluster Oracle Database Free container
- Existing Oracle AI Autonomous Database — Connect to a pre-provisioned Autonomous Database
- Other Existing Database - Connect to another type of pre-existing Oracle AI Database, for example, an Oracle Base DB, or an on-premises Oracle Database
- Multi-Tenant — Run multiple OBaaS instances in one cluster
- Namespace and Scope — Control which namespaces components watch
- SigNoz Existing Secret — Pre-provisioned SigNoz credentials
- SigNoz Cold Storage — Offload older observability data to S3-compatible object storage
- Kafka Enabled — Create a Strimzi-managed Kafka cluster
- Coherence Enabled — Create a Coherence Operator-managed cluster
- Private Registry — Air-gapped and corporate registry setups
- Combining Examples — Layer multiple value files together
- Uninstallation — Teardown instructions
- Next Steps — Deploy applications and configure observability
Quick Start Decision Guide
Choose the example values file that best matches your deployment scenario:
values-default.yaml- Minimal installation with default settingsvalues-sidb-free.yaml- Development or testing with an in-cluster Oracle Database Free containervalues-existing-adb.yaml- Production deployments using an existing OCI Autonomous Databasevalues-byodb.yaml- Production deployments using an existing non-Autonomous Oracle Databasevalues-tenant1.yamlandvalues-tenant2.yaml- Multi-tenant deployments in a shared clustervalues-namespace-override.yaml- Custom ingress namespace watching behaviorvalues-signoz-existing-secret.yaml- Pre-provisioned SigNoz admin credentialsvalues-signoz-cold-storage.yaml- Long-term observability retention using S3-compatible object storagevalues-kafka.yaml- Strimzi-managed Kafka cluster for Kafka workloads and observabilityvalues-coherence.yaml- Coherence Operator-managed cluster for distributed-cache and data-grid workloadsvalues-private-registry.yaml- Air-gapped or private registry environments
If you are unsure where to start, use values-sidb-free.yaml for evaluation or values-existing-adb.yaml for an external OCI Autonomous Database deployment.
Architecture
The deployment uses a two-chart architecture. The charts are separated because the OBaaS prerequisites chart installs cluster-wide CRDs and operators that may only exist once in a cluster, while the OBaaS chart contains namespace-scoped resources that can be safely installed multiple times.
For configuration options supplied by a dependent chart, see the version-pinned dependent Helm chart references. The references include the upstream README and complete values file for the versions bundled with OBaaS 2.1.1.
obaas-prereqs (cluster-scoped, install once):
- cert-manager - Automatic TLS certificate management
- external-secrets - Integration with external secret stores
- metrics-server - Container resource metrics
- kube-state-metrics - Kubernetes object metrics
- strimzi-kafka-operator - Kafka cluster operator
- coherence-operator - Coherence cluster operator
- oraoperator - Oracle Database Operator for Kubernetes
Attempting to install the obaas-prereqs chart multiple times will cause CRD version conflicts, duplicate operator controllers, and resource contention.
obaas (namespace-scoped, install per tenant):
- Envoy Gateway - Default Gateway API controller for external access
- ingress-nginx - Deprecated namespace-specific ingress controller, disabled by default and available for explicit opt-in
- Apache APISIX - API Gateway
- Eureka - Service discovery
- Signoz - Observability stack with ClickHouse
- Spring Boot Admin - Application monitoring
- OTMM - Transaction manager for microservices, including MicroTX Workflow for service orchestration
- Kafka cluster - Optional namespace-scoped Kafka custom resource managed by the Strimzi operator from
obaas-prereqs - Coherence cluster - Optional namespace-scoped Coherence custom resource managed by the Coherence Operator from
obaas-prereqs
Each instance operates independently in its own namespace with its own gateway resources and observability stack. Deprecated ingress-nginx is installed only when explicitly enabled.
Cluster
├── obaas-system namespace (prerequisites - install once)
│ ├── cert-manager
│ ├── external-secrets
│ ├── metrics-server
│ ├── kube-state-metrics
│ ├── strimzi-kafka-operator (manages Kafka CRs across all namespaces)
│ └── coherence-operator (manages Coherence CRs across all namespaces)
├── tenant1 namespace (OBaaS instance 1)
│ ├── Envoy Gateway resources
│ ├── APISIX, Eureka, and other OBaaS services
│ ├── Kafka cluster (CR managed by Strimzi)
│ ├── Coherence cluster (CR managed by Coherence Operator)
│ └── Signoz + ClickHouse
└── tenant2 namespace (OBaaS instance 2)
├── Envoy Gateway resources
├── APISIX, Eureka, and other OBaaS services
├── Kafka cluster (CR managed by Strimzi)
├── Coherence cluster (CR managed by Coherence Operator)
└── Signoz + ClickHouse
Namespace behavior: All OBaaS chart components deploy to the release namespace (specified with the -n flag during installation). Envoy Gateway is enabled by default. If deprecated ingress-nginx is explicitly enabled, it watches only its own release namespace by default (scope.enabled: true).
Network isolation: The OBaaS chart installs NetworkPolicy resources in the release namespace. Effective enforcement depends on the cluster CNI plugin; use a CNI that supports Kubernetes NetworkPolicy and has policy enforcement enabled. By default, the chart establishes a default-deny baseline, allows same-namespace traffic, allows DNS egress, permits public ingress to the configured gateway or ingress entrypoints, and explicitly allows external egress for compatibility with external databases, OCI APIs, registries, and identity providers.
Directory structure:
helm/infra-charts/
├── obaas-prereqs/ # Cluster-singleton prerequisites (install once per cluster)
└── obaas/ # OBaaS application chart (install N times in different namespaces)
└── examples/ # Example values files for different scenarios
Application & Privileged Database User
During installation, the Helm chart sets up two database users with different privilege levels:
| User | Purpose | Created by |
|---|---|---|
| SYSTEM (non-ADB) or ADMIN (ADB) | Privileged user for one-time database initialization | Pre-exists (external DBs) or auto-generated (container DBs) |
| OBAAS_USER | Application-level user for runtime platform operations | Created by the init script during first install |
OBAAS_USER is the unprivileged database identity used by OBaaS components at runtime. The privileged user is only used once — to create OBAAS_USER and grant it the minimum permissions the platform needs.
What OBAAS_USER is granted
The init script grants or refreshes the following permissions:
- DB_DEVELOPER_ROLE (Oracle 21c+). On Oracle 19c or databases where this role is unavailable, falls back to explicit schema privileges:
CREATE SESSION,CREATE TABLE,CREATE VIEW,CREATE SEQUENCE,CREATE PROCEDURE,CREATE TRIGGER, andCREATE TYPE. - SELECT on monitoring views Required by the Oracle Database Exporter for Prometheus metrics.
- Unlimited quota on its default tablespace.
- ADB only:
EXECUTE ON DBMS_CLOUD_AIandEXECUTE ON DBMS_CLOUD_PIPELINEfor the AI Optimizer feature.
Which components use OBAAS_USER
The following components read the <release>-db-authn Secret at runtime:
- Oracle Database Exporter — connects as OBAAS_USER to query the monitoring views for Prometheus metrics.
- Database init container — uses the credentials during initialization to create the user and verify connectivity.
- MicroTx Workflow Server — connects as OBAAS_USER and uses Flyway to create or update its workflow schema objects when
otmm.workflowServer.enabledistrue.
Other platform components such as Eureka and Admin Server connect to the database through their own Spring Boot datasource configuration.
To use a pre-existing database user instead of the auto-generated OBAAS_USER, create a Secret with your credentials and reference it via database.authN.secretName. The init script will skip user creation if the named user already exists in DBA_USERS.
Prerequisites
Ensure all prerequisites are met before starting the deployment.
Verify that you have the required tools installed:
# Verify Helm installation
helm version
# Verify kubectl and cluster access
kubectl get nodes
Names and Namespace Conventions
The release names obaas and obaas-prereqs, along with the namespace names obaas and obaas-system, are used throughout this document as examples only. Replace them with the release names and namespaces that match your environment.
obaas- example Helm release name for the OBaaS application chartobaas-prereqs- example Helm release name for the prerequisites chartobaas- example namespace for namespace-scoped OBaaS application componentsobaas-system- example namespace for the cluster-scoped prerequisites chart
All namespace-scoped OBaaS resources are deployed into the namespace provided with the -n flag in the Helm and kubectl commands, and the prerequisites chart is installed once per cluster into the namespace you choose for those cluster-level resources.
For example:
helm upgrade --install <prereqs-release> obaas/obaas-prereqs -n <platform-system-namespace> --create-namespace [--debug]
helm upgrade --install <app-release> obaas/obaas -n <application-namespace> --create-namespace [--debug]
kubectl get pods -n <platform-system-namespace>
kubectl get pods -n <application-namespace>
If you are deploying multiple OBaaS instances in the same cluster, each instance should use its own namespace. The prerequisite chart should still be installed once per cluster in a single shared namespace of your choice.
Installation Guide
Step 0: Configure Helm Repository
Add the Helm repository to your local Helm installation using the following commands:
helm repo add obaas https://oracle.github.io/microservices-backend/helm
helm repo update
Step 1: Install cert-manager
OBaaS requires cert-manager as a prerequisite. If you do not have cert-manager installed on your cluster, install it now:
helm install \
cert-manager oci://quay.io/jetstack/charts/cert-manager \
--version v1.21.1 \
--namespace cert-manager \
--create-namespace \
--set crds.enabled=true \
--set crds.keep=false
Step 2: Install Prerequisites (Once Per Cluster)
Only install prerequisites once per cluster. Installing multiple times will cause CRD conflicts and duplicate operator controllers.
helm upgrade --install <prereqs-release> obaas/obaas-prereqs -n <platform-system-namespace> --create-namespace [--debug] [--values <path_to_custom_values>]
Verify all prerequisite pods are running before proceeding:
kubectl get pods -n <platform-system-namespace>
All pods should reach Running status within 2-3 minutes.
To verify that Metrics Server is serving resource metrics, check its aggregated APIService and query node or pod usage:
kubectl get apiservice v1beta1.metrics.k8s.io
kubectl top nodes
kubectl top pods -n <application-namespace>
The APIService should report Available=True. A temporarily empty kubectl top result can be normal immediately after startup while the first scrape
completes.
Step 3: Choose a Values File and Install OBaaS
Choose an example configuration that matches your deployment scenario and install:
helm upgrade --install <app-release> obaas/obaas -f examples/<values-file>.yaml -n <application-namespace> --create-namespace [--debug]
See Example Configurations below for the full list of available examples.
Monitor the installation:
kubectl get pods -n <application-namespace> -w
It may take 5-10 additional minutes for all pods to reach the Running state.
Step 4: Verify Installation
After the installation completes, verify all components are running:
kubectl get pods -A
Expected results:
- All pods in
Runningstate - Services with ClusterIP or LoadBalancer IPs assigned
- Ingress configured with your hostname
Example Configurations
Several example configurations are provided for comparison.
| Values file | Best for | External DB required | Notes |
|---|---|---|---|
values-default.yaml | Minimal evaluation | No | Uses chart defaults |
values-sidb-free.yaml | Development and testing | No | Runs Oracle Database Free in the cluster |
values-existing-adb.yaml | OCI production | Yes | Uses Autonomous Database |
values-byodb.yaml | Existing Oracle Database | Yes | Non-Autonomous Oracle Database only |
values-tenant1.yaml, values-tenant2.yaml | Multi-tenant setups | Depends | Requires unique ingress settings per tenant |
values-namespace-override.yaml | Namespace watch tuning | Depends | Adjusts ingress scope |
values-signoz-existing-secret.yaml | GitOps and pre-created credentials | Depends | Uses an existing SigNoz secret |
values-signoz-cold-storage.yaml | Long-term observability retention | Depends | Uses S3-compatible object storage |
values-kafka.yaml | Kafka workloads and observability testing | Depends | Creates a Strimzi-managed Kafka cluster |
values-coherence.yaml | Coherence integration and distributed-cache testing | Depends | Creates a Coherence Operator-managed cluster |
values-private-registry.yaml | Air-gapped and private registry installs | Depends | Mirrors images to a private registry |
The OBaaS chart selects the database mode with database.type. The supported values are SIDB-FREE, ADB-FREE, ADB-S, and OTHER; choose the example values file that matches the database deployment you plan to use.
Default Configuration (values-default.yaml)
Minimal configuration with no overrides. All subcharts use their default settings.
Use case: Quick start, development, testing
Installation:
helm upgrade --install <app-release> obaas/obaas -f examples/values-default.yaml -n <application-namespace> --create-namespace [--debug]
SIDB-FREE Database (values-sidb-free.yaml)
If you use SIDB, you may need more ephemeral storage on your nodes. Please refer to prerequisites for details.
Uses Oracle Database Free as an in-cluster container. This is the default database type.
Use case: Development, testing, standalone deployments
Installation:
helm upgrade --install <app-release> obaas/obaas -f examples/values-sidb-free.yaml -n <application-namespace> --create-namespace [--debug]
Existing Oracle AI Autonomous Database Configuration (values-existing-adb.yaml)
Connects to an existing OCI Autonomous Database (ADB-S) instead of deploying a database container.
Use case: Production deployments using a pre-provisioned OCI Oracle Autonomous AI Database
Prerequisites: Create required secrets before installing
-
Create the OCI API key secret.
python3 tools/oci_config.py --namespace <application-namespace> [--config <config-file>] [--profile <profile-name>]<config-file>is the location of your OCI configuration file, e.g.,/home/user/.oci/config<profile>is the profile in your config file to use, if notDEFAULTnotePython 3.12 or later is required to run the
oci_config.pyscript. -
Create the privileged authentication secret for the
ADMINuser. Replace<db-name>with the name of your database.kubectl -n <application-namespace> create secret generic <db-name>-db-priv-authn \--from-literal=username=ADMIN \--from-literal=password=<admin-password> \--from-literal=service=<db-name>_tp -
Optional: create the application user secret for
OBAAS_USER. Replace<db-name>with the name of your database.kubectl -n <application-namespace> create secret generic <db-name>-db-authn \--from-literal=username=OBAAS_USER \--from-literal=password=<app-user-password> \--from-literal=service=<db-name>_tp
Get the OCID for an Existing Oracle AI Autonomous Database
Before installing OBaaS with an existing Oracle AI Autonomous Database (ADB), you must get the database OCID. The OCID is required as an input parameter for the helm installation command.
A helper script, get-adb-ocid.sh, is provided to retrieve the OCID of an existing Oracle AI Autonomous Database. Alternatively, you can locate the OCID manually in the Oracle Cloud Infrastructure (OCI) Console.
Script Usage
tools/get-adb-ocid.sh -r <region> (-c <compartment-name> | --compartment-ocid <ocid>) -dbname <adb-display-name> [options]
Parameters
| Parameter | Description |
|---|---|
-r <region> | OCI region where the Oracle AI Autonomous Database is deployed |
-c <compartment-name> | Name of the OCI compartment containing the database |
--compartment-ocid <ocid> | OCID of the OCI compartment containing the database |
-dbname <adb-display-name> | Display name of the existing Oracle AI Autonomous Database |
[options] | Additional optional parameters supported by the script |
Example
tools/get-adb-ocid.sh \
-r us-ashburn-1 \
-c my-compartment \
-dbname my-existing-adb
The script returns the OCID for the specified Autonomous Database, which can then be used during the OBaaS installation process.
Installation:
The --set database.authN.secretName=<app-user-secret> argument is optional.
helm upgrade --install <app-release> obaas/obaas \
-n <application-namespace> \
-f examples/values-existing-adb.yaml \
--set database.oci.ocid=<adb-ocid> \
--set database.privAuthN.secretName=<admin-user-secret> \
--set database.authN.secretName=<app-user-secret> \
[--debug]
Other Existing Oracle AI Database (values-byodb.yaml)
Connects to an existing Oracle AI Database using a connection string and user credentials. Do not use this option for an Oracle Autonomous Database. This is a good option for an Oracle Base DB or an on-premises Oracle AI Database.
Use case: Production deployments using a pre-existing Oracle AI Database (non-Autonomous)
Prerequisites: Create required secrets before installing
-
Create the privileged authentication secret for an appropriate admin user.
kubectl -n <application-namespace> create secret generic <db-name>-db-priv-authn \--from-literal=username=SYSTEM \--from-literal=password=<system-password> \--from-literal=service=<service-name> -
Use an admin user with DBA privileges that can create application users and grant the required permissions. The
SYSTEMuser is a good choice. This user should not require theSYSDBArole. -
Ensure the admin user has the following permissions:
SELECT WITH GRANT OPTION on:
DBA_TABLESPACE_USAGE_METRICS, DBA_TABLESPACES,
GV_$SYSTEM_WAIT_CLASS, GV_$ASM_DISKGROUP_STAT, GV_$DATAFILE,
GV_$SYSSTAT, GV_$PROCESS, GV_$WAITCLASSMETRIC, GV_$SESSION,
GV_$RESOURCE_LIMIT, GV_$PARAMETER, GV_$DATABASE,
GV_$SQLSTATS, GV_$SYSMETRIC, GV_$CON_SYSMETRIC, V_$DIAG_ALERT_EXT
EXECUTE WITH GRANT OPTION on:
SYS.DBMS_AQ, SYS.DBMS_AQADM, SYS.DBMS_AQIN,
SYS.DBMS_AQIN, SYS.DBMS_AQJMS_INTERNAL
- Review and update the database connection details in
examples/values-byodb.yaml.
Installation:
helm upgrade --install <app-release> obaas/obaas \
-n <application-namespace> \
-f examples/values-byodb.yaml \
[--debug]
Multi-Tenant Setup (values-tenant1.yaml, values-tenant2.yaml)
Configures unique IngressClass and credentials for each tenant to allow multiple OBaaS instances in the same cluster.
Use case: Multi-tenant deployments, namespace isolation
Each tenant must have unique values for the following to avoid conflicts:
controller.ingressClasscontroller.ingressClassResource.namecontroller.ingressClassResource.controllerValuecontroller.electionID
Installation:
# Install first tenant
helm upgrade --install <tenant1-release> obaas/obaas -n <tenant1-namespace> -f examples/values-tenant1.yaml --create-namespace [--debug]
# Install second tenant
helm upgrade --install <tenant2-release> obaas/obaas -n <tenant2-namespace> -f examples/values-tenant2.yaml --create-namespace [--debug]
Namespace and Scope Configuration (values-namespace-override.yaml)
Demonstrates legacy opt-in ingress-nginx watching scope configuration. All OBaaS components deploy to the release namespace (specified with the -n flag).
Use case: Preserving legacy ingress-nginx namespace watch settings when ingress-nginx is explicitly enabled
Installation:
helm upgrade --install <app-release> obaas/obaas -f examples/values-namespace-override.yaml -n <application-namespace> --create-namespace [--debug]
SigNoz Existing Secret (values-signoz-existing-secret.yaml)
Uses a pre-existing Kubernetes secret for SigNoz admin authentication instead of auto-generating one.
Use case: GitOps workflows, pre-provisioned credentials
Prerequisites:
- Create the secret before installing the chart.
Installation:
# Create the secret first
kubectl create secret generic my-signoz-secret \
--from-literal=email=admin@mydomain.com \
--from-literal=password=my-secure-password \
-n <application-namespace>
# Install with the example
helm upgrade --install <app-release> obaas/obaas -f examples/values-signoz-existing-secret.yaml -n <application-namespace> [--debug]
The setup Job discovers the organization associated with the Secret's email
through the SigNoz API. If that account belongs to multiple organizations, include
an orgID key containing the selected organization's UUID in the same Secret.
See dashboard setup for
reconciliation behavior and verification commands.
Upgrade SigNoZ during an existing-release upgrade
Follow the OBaaS upgrade guide for prerequisite and Envoy Gateway CRD updates and the two-step ClickHouse/SigNoZ upgrade. Older ClickHouse versions must be upgraded before the new pre-upgrade migration runs.
The two-step procedure without protected recovery is for installations where
observability data loss is acceptable. For backup and preservation, follow
SigNoZ's ClickHouse upgrade and backup guidance.
Do not set signozUpgrade.mode=destructive-replace for the in-place upgrade.
SigNoz Cold Storage (values-signoz-cold-storage.yaml)
Configures SigNoz ClickHouse cold storage so recent telemetry stays on the local persistent disk and older data is offloaded to S3-compatible object storage.
This example can be used with Oracle Cloud Infrastructure Object Storage or with an on-premises S3-compatible object store such as MinIO.
Use case: Longer telemetry retention, reduced local disk pressure, and support for both cloud and on-premises object storage backends
Prerequisites: Prepare object storage settings before installing
-
Create or identify an S3-compatible bucket for SigNoz cold storage.
-
Collect the connection details for your object store:
endpoint,accessKey, andsecretAccess. -
Review and update
examples/values-signoz-cold-storage.yamlwith your environment-specific values. -
Size local ClickHouse storage appropriately.
signoz.clickhouse.persistence.sizecontrols hot storage capacity, andsignoz.clickhouse.coldStorage.defaultKeepFreeSpaceBytescontrols how much local disk remains free before older data is offloaded. -
For on-premises installations, you can use an S3-compatible endpoint provided by MinIO instead of OCI Object Storage. (min.io)
Installation:
helm upgrade --install <app-release> obaas/obaas \
-f examples/values-signoz-cold-storage.yaml \
-n <application-namespace> --create-namespace [--debug]
Kafka Enabled Configuration (values-kafka.yaml)
Creates a Strimzi-managed Kafka cluster in the OBaaS release namespace. This is useful for Kafka integration testing, CloudBank Helidon producer and consumer workloads, and Kafka observability validation in SigNoz.
Prerequisites:
- Install
obaas-prereqsonce per cluster. - Keep
strimzi-kafka-operatorenabled inobaas-prereqs. - Ensure the Strimzi operator watches the OBaaS release namespace.
Installation:
helm upgrade --install <app-release> obaas/obaas \
-f examples/values-kafka.yaml \
-n <application-namespace> \
--create-namespace [--debug]
Optional Kafka metrics in SigNoz:
helm upgrade --install <app-release> obaas/obaas \
-f examples/values-kafka.yaml \
-f extensions/kafka-metrics.yaml \
-n <application-namespace> \
--create-namespace [--debug]
With release name obaas, Kafka clients can use obaas-kafka-cluster-kafka-bootstrap:9092. The chart also creates the stable alias kafka-bootstrap:9092.
Coherence Enabled Configuration (values-coherence.yaml)
Creates a three-member, ephemeral Coherence cluster in the OBaaS release namespace. This is useful for Coherence integration testing and development workloads that require a distributed cache or data grid.
Prerequisites:
- Install
obaas-prereqsonce per cluster. - Keep
coherence-operatorenabled inobaas-prereqs. - Ensure the Coherence Operator watches the OBaaS release namespace.
Installation:
helm upgrade --install <app-release> obaas/obaas \
-f examples/values-coherence.yaml \
-n <application-namespace> \
--create-namespace [--debug]
Set coherence.name to a DNS-compatible name unique within the release namespace. The provided example uses mysample-cluster; replace it with the name for your cluster. Verify the resource and its members with:
kubectl get coherence <coherence-cluster-name> -n <application-namespace>
kubectl get pods -n <application-namespace>
The example does not create persistent storage. Configure coherence.coherence.persistence with a suitable StorageClass and volume size when durable Coherence data is required.
MicroTx Coordinator Database Storage
To persist coordinator transaction state in the OBaaS database, configure:
otmm:
coordinator:
storage:
type: db
With storage.db.credentialSecretName empty, the coordinator reuses the application
credential secret and database wallet used by the metrics exporter, including
custom database.authN secret names and keys. At startup, Python 3 in the
coordinator image serializes the username and password into MicroTx's JSON
credential format, then starts /app/tcs. The default 26.1.2 image contains this
runtime; custom images must provide it too. Coordinator settings are loaded through
envFrom from the chart's ConfigMap, allowing the secret-backed connection string
to take effect. The connection string defaults to the
application secret's service key; set storage.db.connectionString to override it.
For the shared wallet, the chart overlays sqlnet.ora with an explicit wallet
directory of /etc/config/dbwallet. This avoids the downloaded wallet's
?/network/admin path, which does not point to the coordinator's mount. The wallet
Secret itself is unchanged. Custom wallet secrets and ConfigMaps must supply a
sqlnet.ora that matches this mount path.
If an existing values file sets storage.db.credentialSecretName to
microtx-coordinator-db, remove that override or set it to "" to use the shared
application secret. A nonempty value still selects a pre-existing, separate secret
whose storage.db.credentialSecretKey (default secret) contains
{"username":"...","password":"..."}; this mode requires an explicit connection
string. Custom wallet secrets or ConfigMaps can also be selected under storage.db.
Database persistence requires MicroTx Enterprise Edition, which enables durable storage and
the database schema and privileges required by that edition. For storage.type: db,
the chart runs the coordinator image in INIT_CONTAINER_ENABLED=true mode before
starting the service, using the same credentials and wallet. This initializes or
upgrades the MicroTx schema; the user needs its documented schema-creation
privileges and tablespace quota. Check initialize-coordinator-schema container
logs if initialization fails. Set storage.db.initializeSchema: false when a DBA
provisions and upgrades these objects separately. Review the MicroTx migration and
backup requirements before upgrading an existing transaction store. The chart
defaults to coordinator memory storage.
Private Registry Configuration (values-private-registry.yaml)
Uses a private container registry for all images with authentication.
Use case: Air-gapped environments, corporate registries, security compliance
Prerequisites
-
Mirror all required images to your private registry.
A utility is provided to mirror the images to your registry. Provide the name of your target registry. You must be authenticated to the target registry before running this script.
cd helm/infra-charts/tools./mirror-images.sh myregistry.example.com -
Create an image pull secret.
kubectl create secret docker-registry myregistry-secret \--docker-server=myregistry.example.com \--docker-username=<username> \--docker-password=<password> \--docker-email=<email> -
Install both the
obaas-prereqsandobaascharts using values files with your registry details.
Installation:
helm upgrade --install <prereqs-release> obaas/obaas-prereqs -n <platform-system-namespace> -f obaas-prereqs/examples/values-private-registry.yaml --create-namespace [--debug]
helm upgrade --install <app-release> obaas/obaas -n <application-namespace> -f obaas/examples/values-private-registry.yaml --create-namespace [--debug]
Image registry override details
Each subchart has its own image configuration that must be set explicitly.
Subcharts with a dedicated registry field (set to your registry URL):
- ingress-nginx:
controller.image.registrywhen deprecated ingress-nginx is explicitly enabled - signoz:
global.imageRegistry - SigNoz setup Job:
signoz.setup.image.registry, with separatesignoz.setup.image.repositoryandsignoz.setup.image.tagvalues. The default iscontainer-registry.oracle.com/os/oraclelinux:9; this Job does not usesignoz.global.imageRegistry. The private registry example mirrors it asmyregistry.example.com/os/oraclelinux:9. Replacement images must providepython3and its standard library; no packages are installed at runtime. The formersignoz.utilities.curlJq.imageoverride is no longer used. - apisix etcd:
apisix.etcd.image.registry
OBaaS components requiring the full repository path (include registry in image.repository):
- eureka, admin-server, otmm, oracle-database-exporter
- database (for SIDB-FREE/ADB-FREE)
- apisix:
apisix.image.repository,apisix.initContainer.image
Image pull secrets are propagated to subcharts via global.imagePullSecrets. See values-private-registry.yaml for a complete example.
Cluster-singleton prerequisites (cert-manager, external-secrets, metrics-server, kube-state-metrics, strimzi-kafka-operator) have their own image configuration in the obaas-prereqs chart.
Combining Examples
You can layer multiple value files to combine configurations:
# Combine multi-tenant and private registry configurations
helm upgrade --install <tenant1-release> obaas/obaas \
-f examples/values-tenant1.yaml \
-f examples/values-private-registry.yaml \
-n <tenant1-namespace> --create-namespace [--debug]
Or create a custom values file:
# my-custom-values.yaml
global:
imagePullSecrets:
- myregistry-secret
# Disable components you don't need
ai-optimizer:
enabled: false
signoz:
enabled: false
Uninstallation
To remove OBaaS from your cluster:
Uninstalling will delete all resources. Ensure you have backups of any important data before proceeding.
Step 1: Uninstall OBaaS instances first:
helm uninstall <app-release> -n <application-namespace>
helm uninstall <tenant1-release> -n <tenant1-namespace>
helm uninstall <tenant2-release> -n <tenant2-namespace>
Step 2: Delete namespaces (optional):
kubectl delete namespace <application-namespace> <tenant1-namespace> <tenant2-namespace>
Step 3: Uninstall prerequisites (only if removing entirely):
helm uninstall <prereqs-release> -n <platform-system-namespace>
Uninstalling prerequisites will affect all OBaaS instances in the cluster. Only do this if you're removing OBaaS completely.
Next Steps
Once OBaaS is running, you can:
- Deploy your application using the OBaaS platform
- Configure observability for metrics, logs, and traces