qserv-kafka Helm values reference

Helm values reference table for the qserv-kafka application.

Key

Type

Default

Description

config.consumerGroupId

string

"qserv"

Kafka consumer group ID

config.jobCancelTopic

string

"lsst.tap.job-delete"

Kafka topic for query cancellation requests

config.jobRunTopic

string

"lsst.tap.job-run"

Kafka topic for query execution requests

config.jobStatusTopic

string

"lsst.tap.job-status"

Kafka topic for query status

config.logLevel

string

"INFO"

Logging level

config.logProfile

string

"production"

Logging profile (production for JSON, development for human-friendly)

config.maxWorkerJobs

int

2

Maximum number of arq jobs each worker can process simultaneously

config.qservDatabaseOverflow

int

50

Extra database connections that may be opened in excess of the pool size to handle surges in load. This is used primarily by the frontend for jobs that complete immediately.

config.qservDatabasePoolSize

int

2

Database pool size. This is the number of MySQL connections that will be held open regardless of load. This should generally be set to the same as maxWorkerJobs.

config.qservDatabaseUrl

string

None, must be set

URL to the Qserv MySQL interface (must use a scheme of mysql+asyncmy)

config.qservPollInterval

string

"1s"

Interval at which Qserv is polled for query status in Safir parse_timedelta format

config.qservRestMaxConnections

int

20

Maximum simultaneous connections to open to the REST API.

config.qservRestTimeout

string

"30s"

Timeout for REST API calls in Safir parse_timedelta format. This includes time spent waiting for a connection if the maximum number of connections has been reached.

config.qservRestUrl

string

None, must be set

URL to the Qserv REST API

config.resultTimeout

int

3600 (1 hour)

How long to wait for result processing (retrieval and upload) before timing out, in seconds. This doubles as the timeout forcibly terminating result worker pods.

frontend.affinity

object

{}

Affinity rules for the qserv-kafka frontend pod

frontend.nodeSelector

object

{}

Node selection rules for the qserv-kafka frontend pod

frontend.podAnnotations

object

{}

Annotations for the qserv-kafka frontend pod

frontend.resources

object

See values.yaml

Resource limits and requests for the qserv-kafka frontend pod

frontend.tolerations

list

[]

Tolerations for the qserv-kafka frontend pod

global.baseUrl

string

Set by Argo CD

Base URL for the environment

global.host

string

Set by Argo CD

Host name for ingress

global.vaultSecretsPath

string

Set by Argo CD

Base path for Vault secrets

image.pullPolicy

string

"IfNotPresent"

Pull policy for the qserv-kafka image

image.repository

string

"ghcr.io/lsst-sqre/qserv-kafka"

Image to use in the qserv-kafka deployment

image.tag

string

The appVersion of the chart

Tag of image to use

ingress.annotations

object

{}

Additional annotations for the ingress rule

redis.config.secretKey

string

"redis-password"

Key inside secret from which to get the Redis password (do not change)

redis.config.secretName

string

"qserv-kafka"

Name of secret containing Redis password

redis.persistence.accessMode

string

"ReadWriteOnce"

Access mode of storage to request

redis.persistence.enabled

bool

true

Whether to persist Redis storage. Setting this to false will use emptyDir and lose track of all queries on restart. Only use this for a test deployment.

redis.persistence.size

string

"100Mi"

Amount of persistent storage to request

redis.persistence.storageClass

string

nil

Class of storage to request

redis.persistence.volumeClaimName

string

nil

Use an existing PVC, not dynamic provisioning. If this is set, the size, storageClass, and accessMode settings are ignored.

redis.resources

object

See values.yaml

Resource limits and requests for the Redis pod

resultWorker.affinity

object

{}

Affinity rules for the qserv-kafka worker pods

resultWorker.autoscaling.enabled

bool

true

Enable autoscaling of qserv-kafka result workers

resultWorker.autoscaling.maxReplicas

int

10

Maximum number of qserv-kafka worker pods. Each replica will open database connections up to the configured pool size and overflow limits, so make sure the combined connections are under the postgres connection limit.

resultWorker.autoscaling.minReplicas

int

1

Minimum number of qserv-kafka worker pods

resultWorker.autoscaling.targetCPUUtilizationPercentage

int

75

Target CPU utilization of qserv-kafka worker pods.

resultWorker.nodeSelector

object

{}

Node selection rules for the qserv-kafka worker pods

resultWorker.podAnnotations

object

{}

Annotations for the qserv-kafka worker pods

resultWorker.replicaCount

int

1

Number of result worker pods to start if autoscaling is disabled

resultWorker.resources

object

See values.yaml

Resource limits and requests for the qserv-kafka worker pods

resultWorker.tolerations

list

[]

Tolerations for the qserv-kafka worker pods