values.yaml

# values ref: https://github.com/yuvipanda/cryptnono/blob/main/cryptnono/values.yaml
cryptnono:
  enabled: true
  detectors:
    execwhacker:
      metrics:
        enabled: true
      # execwhacker needs access to containerd, and docker on build hosts, to lookup the container.
      # Override if K8s uses a different path for containerd
      containerdHostPath: /run/containerd/containerd.sock
      dockerHostPath: /run/dind/docker.sock

imagePullSecrets:

tags: {}

etcJupyter:
  jupyter_notebook_config.json:
    NotebookApp:
      allow_origin: "*"
      tornado_settings:
        trust_xheaders: true
      # shutdown the server after no activity
      shutdown_no_activity_timeout: 600

    # if a user leaves a notebook with a running kernel,
    # the effective idle timeout will typically be CULL_TIMEOUT + CULL_KERNEL_TIMEOUT
    # as culling the kernel will register activity,
    # resetting the no_activity timer for the server as a whole
    MappingKernelManager:
      # shutdown kernels after no activity
      cull_idle_timeout: 600
      # check for idle kernels this often
      cull_interval: 60
      # a kernel with open connections but no activity still counts as idle
      # this is what allows us to shutdown servers
      # when people leave a notebook open and wander off
      cull_connected: true

# values ref: https://github.com/jupyterhub/binderhub/blob/main/helm-chart/binderhub/values.yaml
# can't use binderhub.enabled due to validation errors
binderhubEnabled: true
binderhub:
  replicas: 2

  resources:
    requests:
      cpu: "0.25"
      memory: 1Gi
    limits:
      cpu: "2"
      memory: 1Gi

  hpa:
    enabled: false
    minReplicas: 1
    maxReplicas: 1
    targetCPU: 100 # this is in percent of requests.cpu

  networkPolicy:
    enabled: true
    egress:
      tcpPorts:
        - 80 # http
        - 443 # https
        - 9418 # git
        - 873 # rsync
        - 1094 # xroot
        - 1095 # xroot
        - 16286 # Wolfram Engine on-demand licensing
        - 4001 # IPFS
      cidr: 0.0.0.0/0
    ingress:
      bannedIps: []

  config:
    GitHubRepoProvider:
      # Add banned repositories to the list below
      # They should be strings that will match "^<org-name>/<repo-name>.*"
      banned_specs:
        # e.g. '^org/repo.*'
        - .*xmrig.*
        - ^a2nk/.*
        - ^imhajes/.*
        - ^ines/spacy-binder.*
        - ^soft4voip/rak.*
        - ^hmharshit/cn-ait.*
        - ^shishirchoudharygic/mltraining.*
        - ^hmharshit/mltraining.*
        - ^FDesnoyer/MathExp.*
        - ^GuitarsAI/.*
        # ferarussia is clearly a fake GitHub account created by GuitarsAI to get around ban
        # it was created the day GuitarsAI was blocked and does the same thing
        - ^ferarussia/.*
        - ^irw4ns1/.*
        # CryptoCurrency mining bans
        - ^ajiBal/.*
        - ^wyattsanchez54/.*
        - ^walterpowell201/.*
      high_quota_specs:
        []
        # - ^jupyterlab/.*
        # - ^jupyter/.*
        # - ^jupyterhub/.*
        # - ^jupyter-widgets/.*
      spec_config:
        - pattern: ^ipython/ipython-in-depth.*
          config:
            quota: 100
        - pattern: ^petlja/.*
          config:
            quota: 50
        # - pattern: ^github-owner/github-repo-prefix.*
        #   # YYYY-MM-DD of workshop
        #   config:
        #     quota: 123
        - pattern: ^MNGuenther/hackanexoplanet-env.*
          # 2023-04-18
          # https://github.com/jupyterhub/mybinder.org-deploy/issues/2534
          config:
            quota: 300

    GitRepoProvider:
      banned_specs:
        - ^(git|https?)%3A%2F%2Fgithub.com%2Fa2nk%2F.*
        - ^https%3A%2F%2Fbitbucket.org%2Fnikiubel%2Fnikiubel.bitbucket.io.git/.*
        - ^https%3A%2F%2Fjovian.ml%2Fapi%2Fgit%2F.*
        - ^https%3A%2F%2Fframagit.org%2FCecGhesq%2Flic_mdf_nsi_1.*
        - ^(git|https?)%3A%2F%2Fnotabug.org%2FulslcuRux3Y%2F.*
        - ^(git|https?)%3A%2F%2Fgitlab.com%2Fjasmt507%2F.*
        - ^(git|https?)%3A%2F%2Fgitlab.com%2Fh4j3s1978%2F.*
        - .*%2Fabooz.*
        - .*%2Firw4ns1.*

    GitLabRepoProvider:
      banned_specs:
        - ^h4j3s1978%2F.*
        - ^jasmt507%2F.*
        - .*%2Fabooz.*
      spec_config: []

    BinderHub:
      use_registry: true
      per_repo_quota: 100
      per_repo_quota_higher: 200
      cors_allow_origin: "*"

      banner_message: |
        <a class="btn" style="width:fit-content;height:fit-content;padding:10px;background-color:#e66581;color:white;font-weight:bold;position:absolute;top:0px;bottom:0px;margin:auto 0;right:4px;"
          onmouseover="this.style.backgroundColor='#d15b75'" onmouseout="this.style.backgroundColor='#e66581'"
          href="https://jupyter.org/about#donate" target="_blank">
            🤍 Donate to mybinder.org!
        </a>
        <div style="text-align:center;">
        Thanks to <a href="https://www.ovh.com/">OVH</a>, <a href="https://notebooks.gesis.org">GESIS Notebooks</a> and <a href="https://curvenote.com">Curvenote</a> for supporting us! 🎉
        </div>
        <div style="text-align:center;">
        mybinder.org has updated the base image to Ubuntu 22.04! See the <a href="https://repo2docker.readthedocs.io/en/latest/howto/breaking_changes.html">upgrade guide</a> for details.
        </div>
      about_message: |
        <p>mybinder.org is public infrastructure operated by the <a href="https://jupyterhub-team-compass.readthedocs.io/en/latest/team.html#binder-team">Binder Project team</a>.<br /><br />
        The Binder Project is a member of <a href="https://jupyter.org">Project Jupyter</a>.
        Donations are managed by <a href="https://lf-charities.org">LF Charities</a>, a US 501c3 non-profit.<br /><br />
        For abuse please email: <a href="mailto:binder-team@googlegroups.com">binder-team@googlegroups.com</a>, to report a
        security vulnerability please see: <a href="https://mybinder.readthedocs.io/en/latest/faq.html#where-can-i-report-a-security-issue">Where can I report a security issue</a><br /><br />
        For more information about the Binder Project, see <a href="https://mybinder.readthedocs.io/en/latest/about.html">the About Binder page</a></p>

      extra_footer_scripts:
        01-matomo: |
          // Only load Matomo if DNT is not set.
          // This respects user preferences, and gives us a full score on uBlock origin
          if (navigator.doNotTrack != "1" && // Most Firefox & Chrome
            window.doNotTrack != "1" && // IE & Safari
            navigator.msDoNotTrack != "1" // Old IE
          ) {
          console.log("Loading Matomo tracking, since Do Not Track is off");
            var _paq = _paq || [];
            /* tracker methods like "setCustomDimension" should be called before "trackPageView" */
            // this is matomo's own respect-DoNotTrack
            // should be redundant, but good to be extra explicit
            _paq.push(["setDoNotTrack", true]);
            // disable tracking cookies
            _paq.push(["disableCookies"]);
            _paq.push(['trackPageView']);
            (function() {
              var u="//" + window.location.hostname + "/matomo/";
              _paq.push(['setTrackerUrl', u+'piwik.php']);
              _paq.push(['setSiteId', '1']);
              var d=document, g=d.createElement('script'), s=d.getElementsByTagName('script')[0];
              g.type='text/javascript'; g.async=true; g.defer=true; g.src=u+'piwik.js'; s.parentNode.insertBefore(g,s);
            })();
          }
    KubernetesBuildExecutor:
      build_image: quay.io/jupyterhub/repo2docker:2024.07.0-68.gdd097a2
      memory_limit: "3G"
      memory_request: "1G"

  extraConfig:
    # Send Events to StackDriver on Google Cloud
    # This doesn't need any extra permissions, since the GKE nodes have
    # permission to write to StackDriver by default. We don't block access
    # to cloud metadata in binderhub pod, so this should 'just work'.
    01-eventlog: |
      import os

      import google.cloud.logging
      import google.cloud.logging.handlers
      from traitlets.log import get_logger

      # importing google cloud configures a root log handler,
      # which prevents tornado's pretty-logging
      import logging
      logging.getLogger().handlers = []

      class JSONCloudLoggingHandler(google.cloud.logging.handlers.CloudLoggingHandler):
          def emit(self, record):
              record.name = None
              super().emit(record)

      def _make_eventsink_handler(el):
          client = google.cloud.logging.Client()
          # These events are not parsed as JSON in stackdriver, so give it a different name
          # for now. Should be fixed in https://github.com/googleapis/google-cloud-python/pull/6293
          name = os.environ.get("EVENT_LOG_NAME") or "binderhub-events-text"
          get_logger().info("Sending event logs to %s/logs/%s", client.project, name)
          return [JSONCloudLoggingHandler(client, name=name)]
      c.EventLog.handlers_maker = _make_eventsink_handler

    # backport event_loop_seconds metric
    # from https://github.com/jupyterhub/jupyterhub/pull/4615
    02-event-loop-metric: |
      import time

      from prometheus_client import Histogram
      from tornado.ioloop import PeriodicCallback
      from traitlets.log import get_logger

      c = get_config()  # noqa
      EVENT_LOOP_INTERVAL_SECONDS = Histogram(
          'event_loop_interval_seconds',
          'Distribution of measured event loop intervals',
          namespace="jupyterhub",
          # Increase resolution to 5ms below 50ms
          # because this is where we are most sensitive.
          # No need to have buckets below 25, since we only measure every 20ms.
          buckets=[
              # 5ms from 25-50ms
              25e-3,
              30e-3,
              35e-3,
              40e-3,
              45e-3,
              50e-3,
              # from here, default prometheus buckets
              75e-3,
              0.1,
              0.25,
              0.5,
              0.75,
              1,
              2.5,
              5,
              7.5,
              10,
              float("inf"),
          ],
      )


      class EventLoopMetric:
          event_loop_interval_resolution = 20e-3
          event_loop_interval_log_threshold = 1
          def _event_loop_tick(self):
              """Measure a single tick of the event loop
          
              This measures the time since the last tick
              """
              now = time.perf_counter()
              tick_duration = now - self._last_tick
              self._last_tick = now
              EVENT_LOOP_INTERVAL_SECONDS.observe(tick_duration)
              if tick_duration >= self.event_loop_interval_log_threshold:
                  # warn about slow ticks
                  self.log.warning("Event loop was unresponsive for %.2fs!", tick_duration)
              
          
          def start(self):
              self.log = get_logger()
              self.log.info("starting!")
              now = time.perf_counter()
              self._last_tick = self._last_tick_collect = now
              pc = PeriodicCallback(
                  self._event_loop_tick,
                  self.event_loop_interval_resolution * 1000,
              )
              pc.start()

      metric = EventLoopMetric()
      metric.start()

  registry:
    url: https://gcr.io

  service:
    type: ClusterIP

  ingress:
    enabled: true
    annotations:
      kubernetes.io/ingress.class: nginx
    https:
      enabled: true
      type: kube-lego

  imageBuilderType: dind
  dind:
    resources:
      requests:
        cpu: "0.5"
        memory: 1Gi
      limits:
        cpu: "4"
        memory: 4Gi

  imageCleaner:
    enabled: true
    # when 80% of inodes are used,
    # cull images until only 40% are used.
    imageGCThresholdHigh: 80
    imageGCThresholdLow: 40

  jupyterhub:
    cull:
      # cull every 11 minutes so it is out of phase
      # with the proxy check-routes interval of five minutes
      every: 660
      timeout: 600
      # maxAge is 6 hours: 6 * 3600 = 21600
      maxAge: 21600
    hub:
      networkPolicy:
        enabled: true
      resources:
        requests:
          cpu: "0.25"
          memory: 1Gi
        limits:
          cpu: "2"
          memory: 1Gi
      extraConfig:
        neverRestart: |
          c.KubeSpawner.extra_pod_config.update({'restart_policy': 'Never'})
        noPrivilegeEscalation: |
          c.KubeSpawner.allow_privilege_escalation = False
        noAuthMetrics: |
          c.JupyterHub.authenticate_prometheus = False
      config:
        BinderSpawner:
          cors_allow_origin: "*"
        JupyterHub:
          # only serve the hub's API, not full UI
          hub_routespec: "/hub/api/"
        Proxy:
          # default-route to our nginx "Binder not found" service
          extra_routes:
            "/": "http://proxy-patches"
      service:
        annotations:
          prometheus.io/scrape: "true"
          prometheus.io/path: "/hub/metrics"
    proxy:
      service:
        type: ClusterIP
      chp:
        networkPolicy:
          enabled: true
        resources:
          requests:
            memory: 320Mi
            cpu: "0.1"
          limits:
            memory: 320Mi
            cpu: "0.5"
        # FIXME: move to errorTarget/defaultTarget when hub chart dependency is bumped
        # to include https://github.com/jupyterhub/zero-to-jupyterhub-k8s/pull/2079
        # this still works, though, as repeatedly specifying CLI flags overrides earlier values
        extraCommandLineFlags:
          # defaultTarget needs to wait for jupyterhub 1.4
          # https://github.com/jupyterhub/jupyterhub/pull/3373
          # - --default-target=http://$(PROXY_PATCHES_SERVICE_HOST):$(PROXY_PATCHES_SERVICE_PORT)
          - --error-target=http://$(PROXY_PATCHES_SERVICE_HOST):$(PROXY_PATCHES_SERVICE_PORT)/hub/error
          - --log-level=error
    ingress:
      enabled: true
      annotations:
        ingress.kubernetes.io/proxy-body-size: 64m
        kubernetes.io/ingress.class: nginx
        kubernetes.io/tls-acme: "true"
    scheduling:
      userScheduler:
        enabled: true
        replicas: 2
      podPriority:
        enabled: true
      userPlaceholder:
        enabled: true
        # replicas set in config/<deployment>
    singleuser:
      # clear singleuser.defaultUrl config from chart
      defaultUrl:
      cloudMetadata:
        # we do this in our own network policy
        blockWithIptables: false
      networkPolicy:
        enabled: true
        egress: []
        egressAllowRules:
          nonPrivateIPs: false
      memory:
        guarantee: 450M
        limit: 2G
      cpu:
        guarantee: 0.01
        limit: 1
      storage:
        extraVolumes:
          - name: etc-jupyter
            configMap:
              name: user-etc-jupyter
          - name: etc-jupyter-templates
            configMap:
              name: user-etc-jupyter-templates
        extraVolumeMounts:
          - name: etc-jupyter
            mountPath: /etc/jupyter
          - name: etc-jupyter-templates
            mountPath: /etc/jupyter/templates

      initContainers:
        - name: tc-init
          image: jupyterhub/mybinder.org-tc-init:set-by-chartpress
          imagePullPolicy: IfNotPresent
          env:
            - name: WHITELIST_CIDR
              value: 10.0.0.0/8
            - name: EGRESS_BANDWIDTH
              value: 1mbit
          securityContext:
            # capabilities.add seems to be disabled
            # by the `runAsUser: 1000` in the pod-level securityContext
            # unless we explicitly run as root
            runAsUser: 0
            capabilities:
              add:
                - NET_ADMIN

# values ref: https://github.com/kubernetes/ingress-nginx/blob/main/charts/ingress-nginx/values.yaml
ingress-nginx:
  enabled: true
  rbac:
    create: true
  defaultBackend:
    enabled: true
    minAvailable: 0
  controller:
    admissionWebhooks:
      enabled: false
    resources:
      requests:
        cpu: 100m
        memory: 150Mi
      limits:
        cpu: 800m
        memory: 500Mi
    tolerations:
      - key: "node.kubernetes.io/unschedulable"
        operator: "Exists"
        effect: "NoExecute"
        tolerationSeconds: 30
    affinity:
      podAntiAffinity:
        preferredDuringSchedulingIgnoredDuringExecution:
          - weight: 100
            podAffinityTerm:
              labelSelector:
                matchExpressions:
                  - key: app.kubernetes.io/name
                    operator: In
                    values:
                      - ingress
                  - key: app.kubernetes.io/component
                    operator: In
                    values:
                      - controller
              topologyKey: kubernetes.io/hostname
    replicaCount: 3
    scope:
      enabled: true
    config:
      # Allow POSTs of upto 64MB, for large notebook support.
      proxy-body-size: 64m
    stats:
      enabled: true
    metrics:
      enabled: true
      service:
        annotations:
          prometheus.io/scrape: "true"
          prometheus.io/port: "10254"
    service:
      # Preserve client IPs
      externalTrafficPolicy: Local

redirector:
  enabled: false
  nodeSelector: {}
  redirects: []
  ingress:
    annotations:
      kubernetes.io/ingress.class: nginx
      kubernetes.io/tls-acme: "true"
    tls:
      secretName: kubelego-tls-redirector

static:
  ingress:
    annotations:
      kubernetes.io/ingress.class: nginx
      kubernetes.io/tls-acme: "true"
    tls:
      secretName: kubelego-tls-static

# values ref: https://github.com/grafana/helm-charts/blob/main/charts/grafana/values.yaml
grafana:
  enabled: true
  ingress:
    enabled: true
    annotations:
      kubernetes.io/ingress.class: nginx
      kubernetes.io/tls-acme: "true"
  # deploymentStrategy.type is set to Recreate as we have storage that can only
  # be attached once, we can't have two replicas as RollingUpdate leads to.
  deploymentStrategy:
    type: Recreate
  persistence:
    enabled: true
    size: 1Gi
    accessModes:
      - ReadWriteOnce

  grafana.ini:
    auth.anonymous:
      enabled: true
      org_name: Main Org.
      org_role: Viewer
    auth.basic:
      enabled: true
    smtp:
      enabled: true
    security:
      allow_embedding: true

# values ref: https://github.com/prometheus-community/helm-charts/blob/main/charts/prometheus/values.yaml
prometheus:
  enabled: true
  alertmanager:
    enabled: false
  prometheus-pushgateway:
    enabled: false
  server:
    podLabels:
      # needs access to the Hub API
      hub.jupyter.org/network-access-hub: "true"
    strategy:
      # The default of RollingUpdate fail because attached storage can only be
      # mounted on one pod, so we need to use Recreate that first shut down the
      # pod and then starts it up during updates.
      type: Recreate
    ingress:
      enabled: true
      annotations:
        kubernetes.io/ingress.class: nginx
        kubernetes.io/tls-acme: "true"

  # make sure we collect metrics on pods by app/component at least
  kube-state-metrics:
    metricLabelsAllowlist:
      - pods=[app,component]
      - nodes=[*]

proxyPatches:
  enabled: true
  nodeSelector: {}

matomo:
  enabled: false
  nodeSelector: {}
  resources: {}
  replicas: 1
  service:
    type: ClusterIP
  db:
    username: matomo
    name: matomo
    tables_prefix: matomo_

gcsProxy:
  enabled: false
  replicas: 1
  nodeSelector: {}

analyticsPublisher:
  enabled: false
  events:
    logName: binderhub-events-text
  image:
    name: jupyterhub/mybinder.org-analytics-publisher
    tag: "set-by-chartpress"
  cloudCosts:
    # All billing info goes into same bucket in prod, to which staging has access
    enabled: true
    sourceBucket: binder-billing
    fileName: cloud-costs.jsonl
    kind: csv
  nodeSelector: {}
  resources:
    requests:
      cpu: 10m
      memory: 150Mi
    limits:
      cpu: 100m
      memory: 300Mi

# this is defined in secrets/ for the OVH cluster
eventsArchiver:
  serviceAccountKey: ""

federationRedirect:
  enabled: false
  host: mybinder.org
  resources:
    requests:
      cpu: 10m
      memory: 50Mi
    limits:
      cpu: 100m
      memory: 250Mi
  image:
    name: jupyterhub/mybinder.org-federation-redirect
    tag: "set-by-chartpress"
  check:
    period: 15
    jitter: 0.1
    retries: 5
    timeout: 5
    failed_period: 90
  load_balancer: "rendezvous"
  pod_headroom: 10
  hosts: {}
  nodeSelector: {}

minesweeper:
  enabled: true
  image: jupyterhub/mybinder.org-minesweeper:set-by-chartpress

# cluster-autoscaler is responsible for understanding if k8s nodes needs to be
# added or removed. A pending pod is a sign that another node should be added,
# and an underused node is a sign that a node should be removed.
#
# We always need a cluster-autoscaler, but only deploy it ourselves if its not
# provided as part of the k8s cluster.
#
# values ref: https://github.com/kubernetes/autoscaler/blob/master/charts/cluster-autoscaler/values.yaml
#
cluster-autoscaler:
  enabled: false

# Name:Priority pairs of priority classes to create
# https://kubernetes.io/blog/2023/01/12/protect-mission-critical-pods-priorityclass/
priorityClasses: {}

binderhub-container-registry-helper:
  enabled: false

awsEcrRegistryCleaner:
  enabled: false
  image: ghcr.io/manics/aws-ecr-registry-cleaner:0.0.1
  expiresAfterPullDays: 7
  # 12 hours
  loopDelay: 43200
  serviceAccount:
    name: binderhub-ecr-registry-cleaner
    annotations: {}
  priorityClassName: ""