Connector configuration
Every value the connector reads. Helm values win over environment variables, environment variables win over defaults. Nothing is read from the network.
Required
Section titled “Required”| Value | Type | Description |
|---|---|---|
token |
string | Enrollment token from the admin console. Exchanged once for a client certificate, then discarded. |
prometheus.url |
string | In-cluster address of the query API. Must be reachable from the connector pod. |
Optional
Section titled “Optional”| Value | Default | Description |
|---|---|---|
relay.endpoint |
relay.vedavid.dev:443 |
Where the connector dials out. Override only when self-hosting the relay. |
query.timeout |
10s |
Per-query deadline. Exceeded queries return an error to the phone rather than a stale value. |
query.maxSeries |
500 |
Hard cap on series returned per query. Protects Prometheus from a bad panel more than it protects the phone. |
rbac.serviceAccount |
vedavid-connector |
Service account the pod runs as. Access decisions are made by your cluster, not by us. |
log.level |
info |
error, warn, info or debug. Query text is never logged above debug. |
A full values file
Section titled “A full values file”token: ${VEDAVID_TOKEN}
prometheus: url: http://prometheus-operated:9090
query: timeout: 10s maxSeries: 500
resources: requests: { cpu: 50m, memory: 64Mi } limits: { cpu: 200m, memory: 128Mi }