This page lists the main values of the promptglass Helm chart (oci://registry.promptglass.ai/charts/promptglass). Run helm show values oci://registry.promptglass.ai/charts/promptglass --version <v> for the full, commented file for your version.
| Key |
Default |
Description |
global.domain |
"" |
Base hostname for console and API |
global.region |
us |
Region label (us, eu, ca, custom) |
global.imageRegistry |
registry.promptglass.ai/promptglass |
Override for private / air-gapped registries |
global.imagePullSecrets |
[] |
Pull secrets |
global.licenseSecret |
promptglass-license |
Secret with key license |
global.airgapped |
false |
Disable all outbound calls |
global.mode |
active |
active or standby (DR) |
| Key |
Default |
Description |
gateway.replicas |
2 |
Ignored when autoscaling is on |
gateway.autoscaling.enabled |
true |
HPA on CPU (and optional RPS custom metric) |
gateway.autoscaling.minReplicas / maxReplicas |
2 / 10 |
|
gateway.forwardProxy.enabled |
true |
Forward-proxy listener |
gateway.forwardProxy.port |
3128 |
|
gateway.forwardProxy.caSecret |
"" |
TLS secret with signing CA for AI-domain inspection |
gateway.forwardProxy.upstreamProxy |
"" |
Chain to an upstream proxy |
gateway.api.enabled |
true |
OpenAI/Anthropic-compatible API |
gateway.api.host |
"" |
Ingress host for API mode |
gateway.buffer.size |
5Gi |
Local event buffer volume |
gateway.resources |
2 CPU / 4Gi limits |
|
gateway.service.type |
ClusterIP |
Use LoadBalancer for the proxy port |
| Key |
Default |
Description |
detectors.workers |
1 |
Async detector worker replicas |
detectors.mlClassifiers |
true |
Load NER/code classifiers (memory heavy) |
detectors.gpu.enabled |
false |
Schedule classifiers on GPU nodes |
detectors.timeoutMs |
150 |
Inline detection budget before on_error applies |
| Key |
Default |
Description |
auth.oidc.issuer / clientId / clientSecretRef |
— |
Console SSO |
auth.saml.metadataUrl |
— |
Alternative to OIDC |
auth.scim.enabled |
true |
SCIM 2.0 endpoint at /scim/v2 |
| Key |
Default |
Description |
encryption.provider |
local |
aws-kms, azure-keyvault, gcp-kms, vault-transit, pkcs11, local (dev only) |
encryption.keyId |
— |
Key ARN / URI |
encryption.pkcs11.* |
— |
HSM library, slot, key label, PIN secret |
| Key |
Default |
Description |
postgresql.bundled |
true |
Run in-cluster PostgreSQL |
postgresql.external.{host,port,database,existingSecret,sslMode} |
— |
External PostgreSQL 15+ |
clickhouse.bundled |
true |
Run in-cluster ClickHouse |
clickhouse.external.{hosts,cluster,database,existingSecret} |
— |
External ClickHouse 24+ |
clickhouse.persistence.size |
200Gi |
|
objectStorage.bundled |
true |
Run in-cluster SeaweedFS (S3-compatible, TLS 1.3 only) |
objectStorage.s3.{endpoint,bucket,region,existingSecret,useIRSA} |
— |
External S3-compatible storage |
nats.cluster.replicas |
3 |
JetStream cluster size |
| Key |
Default |
Description |
ingress.enabled |
true |
|
ingress.className |
nginx |
|
ingress.annotations |
{} |
|
ingress.tls |
true |
|
telemetry.otlpEndpoint |
"" |
OpenTelemetry collector for traces and metrics |
telemetry.serviceMonitor |
false |
Create Prometheus Operator ServiceMonitors |
networkPolicies.enabled |
true |
Restrictive default NetworkPolicies |
podSecurity.restricted |
true |
Pod Security Standards restricted profile |