Skip to content


on-premises objectstore has been discontinued and is not available for general use

S3 object store

objectstore on NAIS stores files in buckets. On-prem uses rook ceph storage nodes as the storage solution. This is persistent storage and available in all clusters. The preferred solution is to use GCP for both applications and buckets, but it is supported on-prem as well.

The objectstore / S3 bucket access is secured with access keys and secret keys stored in kubernetes.


This feature is only available in on-premises clusters.


This feature does not have backups of the data in the buckets, please move to GCP if this is needed.

How to

Ceph object storage user

On NAIS on-prem you're required to add a ceph object storage user. This is done through nais-yaml. After applying changes to the ceph-users.yaml and applying this to the cluster the rook operator will create the necessary secrets used to access the S3 bucket. To get access create a pull-request in navikt/nais-yaml.

Example yaml for ceph user:


kind: CephObjectStoreUser
  name: example-application
  namespace: rook-ceph
  store: objectstore
  displayName: "example-application"

After a few minutes the rook operator in the cluster will create the secret. The keys necessary can then be fetched and stored in vault for mounting and usage runtime for the application. Contact @Sten.Ivar.R√łkke or other nais team members for keys.

The objectstore service endpoint is http://objectstore.rook-ceph.svc.nais.local. The objectstore can also be reached from other clusters using https://objectstore.nais.{domain name}.


General ceph metrics are available from several dashboards in grafana:

Grafana Ceph Cluster

Grafana Ceph OSD

Code examples

Kotlin objectstore / S3 bucket access:

package no.nav.example

import com.amazonaws.auth.AWSStaticCredentialsProvider
import com.amazonaws.auth.BasicAWSCredentials
import com.amazonaws.client.builder.AwsClientBuilder
import org.slf4j.LoggerFactory

object S3Client {

    private val s3: AmazonS3
    private val log = LoggerFactory.getLogger(javaClass)

    init {
        val ev = EnvVarFactory.envVar
        val credentials = BasicAWSCredentials(ev.s3AccessKey, ev.s3SecretKey)"New Client: (host: " + ev.s3Url + " - " + ev.s3Region + ", accesskey-length: " + ev.s3AccessKey.length + "S3 secret key Length: " + ev.s3SecretKey.length)
        s3 = AmazonS3ClientBuilder.standard()
            .withEndpointConfiguration(AwsClientBuilder.EndpointConfiguration(ev.s3Url, ev.s3Region))

    fun persistToS3(file: File): PutObjectResult {
        return s3.putObject(Const.BUCKET, Const.FILE, file)

    fun loadFromS3(): File {
        val tempFile = createTempFile()
            .download(Const.BUCKET, Const.FILE, tempFile)
        return tempFile

    private fun createBucketIfMissing() {
        val bucketList = s3.listBuckets().filter { b -> == Const.BUCKET }
        if (bucketList.isEmpty()) {
  "Creating new bucket as its missing: " + Const.BUCKET)
        if (!s3.doesObjectExist(Const.BUCKET, Const.FILE)) {
  "Creating empty file for persisting what have been pushed: " + Const.FILE)
            s3.putObject(Const.BUCKET, Const.FILE, "")

    private fun transferManager(): TransferManager {
        return TransferManagerBuilder.standard().withS3Client(s3).build()


Feel free to help us out by adding more examples!