cloud/ocis
1
0
Fork 0
mirror of https://github.com/owncloud/ocis.git synced 2025-04-18 23:44:07 +03:00
Code
ocis/services/proxy
History

README.md

Proxy

The proxy service is an API-Gateway for the ownCloud Infinite Scale microservices. Every HTTP request goes through this service. Authentication, logging and other preprocessing of requests also happens here. Mechanisms like request rate limiting or intrusion prevention are not included in the proxy service and must be setup in front like with an external reverse proxy.

The proxy service is the only service communicating to the outside and needs therefore usual protections against DDOS, Slow Loris or other attack vectors. All other services are not exposed to the outside, but also need protective measures when it comes to distributed setups like when using container orchestration over various physical servers.

Authentication

The following request authentication schemes are implemented:

  • Basic Auth (Only use in development, never in production setups!)
  • OpenID Connect
  • Signed URL
  • Public Share Token

Configuring Routes

The proxy handles routing to all endpoints that ocis offers. The currently availabe default routes can be found in the code. Changing or adding routes can be necessary when writing own ocis extensions.

Due to the complexity when defining routes, these can only be defined in the yaml file but not via environment variables.

For overwriting default routes, use the following yaml example:

policies:
  - name: ocis
    routes:
      - endpoint: /
        service: com.owncloud.web.web
      - endpoint: /dav/
        service: com.owncloud.web.ocdav

For adding additional routes to the default routes use:

additional_policies:
  - name: ocis
    routes:
      - endpoint: /custom/endpoint
        service: com.owncloud.custom.custom

A route has the following configurable parameters:

endpoint: ""       # the url that should be routed
service: ""        # the service the url should be routed to
unprotected: false # with false (default), calling the endpoint requires authorization.
                   # with true, anyone can call the endpoint without authorisation.

Automatic User and Group Provisioning

When using an external OpenID Connect IDP, the proxy can be configured to automatically provision users upon their first login.

Prequisites

A number of prerequisites must be met for automatic user provisioning to work:

  • ownCloud Infinite Scale must be configured to use an external OpenID Connect IDP
  • The graph service must be configured to allow updating users and groups (GRAPH_LDAP_SERVER_WRITE_ENABLED).
  • One of the claim values returned by the IDP as part of the userinfo response or the access token must be unique and stable for the user. I.e. the value must not change for the whole lifetime of the user. This claim is configured via the PROXY_USER_OIDC_CLAIM environment variable (see below). A natural choice would e.g. be the sub claim which is guaranteed to be unique and stable per IDP. If a claim like email or preferred_username is used, you have to ensure that the user's email address or username never changes.

Configuration

To enable automatic user provisioning, the following environment variables must be set for the proxy service:

  • PROXY_AUTOPROVISION_ACCOUNTS
    Set to true to enable automatic user provisioning.
  • PROXY_AUTOPROVISION_CLAIM_USERNAME
    The name of an OIDC claim whose value should be used as the username for the autoprovsioned user in ownCloud Infinite Scale. Defaults to preferred_username. Can also be set to e.g. sub to guarantee a unique and stable username.
  • PROXY_AUTOPROVISION_CLAIM_EMAIL
    The name of an OIDC claim whose value should be used for the mail attribute of the autoprovisioned user in ownCloud Infinite Scale. Defaults to email.
  • PROXY_AUTOPROVISION_CLAIM_DISPLAYNAME
    The name of an OIDC claim whose value should be used for the displayname attribute of the autoprovisioned user in ownCloud Infinite Scale. Defaults to name.
  • PROXY_AUTOPROVISION_CLAIM_GROUPS
    The name of an OIDC claim whose value should be used to maintain a user's group membership. The claim value should contain a list of group names the user should be a member of. Defaults to groups.
  • PROXY_USER_OIDC_CLAIM
    When resolving and authenticated OIDC user, the value of this claims is used to lookup the user in the users service. For auto provisioning setups this usually is the same claims as set via PROXY_AUTOPROVISION_CLAIM_USERNAME.
  • PROXY_USER_CS3_CLAIM
    This is the name of the user attribute in ocis that is used to lookup the user by the value of the PROXY_USER_OIDC_CLAIM. For auto provisioning setups this usually needs to be set to username.

How it Works

When a user logs into ownCloud Infinite Scale for the first time, the proxy checks if that user already exists. This is done by querying the users service for users, where the attribute set in PROXY_USER_CS3_CLAIM matches the value of the OIDC claim configured in PROXY_USER_OIDC_CLAIM.

If the users does not exist, the proxy will create a new user via the graph service using the claim values configured in PROXY_AUTOPROVISION_CLAIM_USERNAME, PROXY_AUTOPROVISION_CLAIM_EMAIL and PROXY_AUTOPROVISION_CLAIM_DISPLAYNAME.

If the user does already exist, the proxy checks if the displayname has changed and updates that accordingly via graph service.

Unless the claim configured via PROXY_AUTOPROVISION_CLAIM_EMAIL is the same as the one set via PROXY_USER_OIDC_CLAIM the proxy will also check if the email address has changed and update that as well.

Next, the proxy will check if the user is a member of the groups configured in PROXY_AUTOPROVISION_CLAIM_GROUPS. It will add the user to the groups listed via the OIDC claim that holds the groups defined in the envvar and removes it from all other groups that he is currently a member of. Groups that do not exist in the external IDP yet will be created. Note: This can be a somewhat costly operation, especially if the user is a member of a large number of groups. If the group memberships of a user are changed in the IDP after the first login, it can take up to 5 minutes until the changes are reflected in Infinite Scale.

Claim Updates

OpenID Connect (OIDC) scopes are used by an application during authentication to authorize access to a user's detail, like name, email or picture information. A scope can also contain among other things groups, roles, and permissions data. Each scope returns a set of attributes, which are called claims. The scopes an application requests, depends on which attributes the application needs. Once the user authorizes the requested scopes, the claims are returned in a token.

These issued JWT tokens are immutable and integrity-protected. Which means, any change in the source requires issuing a new token containing updated claims. On the other hand side, there is no active synchronisation process between the identity provider (IDP) who issues the token and Infinite Scale. The earliest possible time that Infinite Scale will notice changes is, when the current access token has expired and a new access token is issued by the IDP, or the user logs out and relogs in.

NOTES

  • For resource optimisation, Infinite Scale skips any checks and updates on groupmemberships, if the last update happened less than 5min ago.

  • Infinite Scale can't differentiate between a group being renamed in the IDP and users being reassigned to a different group.

  • Infinite Scale does not get aware when a group is being deleted in the IDP, a new claim will not hold any information from the deleted group. Infinite Scale does not track a claim history to compare.

Impacts

For shares or space memberships based on groups, a renamed or deleted group will impact accessing the resource:

  • There is no user notification about the inability accessing the resource.
  • The user will only experience rejected access.
  • This also applies for connected apps like the Desktop, iOS or Android app!

To give access for rejected users on a resource, one with rights to share must update the group information.

Automatic Quota Assignments

It is possible to automatically assign a specific quota to new users depending on their role. To do this, you need to configure a mapping between roles defined by their ID and the quota in bytes. The assignment can only be done via a yaml configuration and not via environment variables. See the following proxy.yaml config snippet for a configuration example.

role_quotas:
    <role ID1>: <quota1>
    <role ID2>: <quota2>

Automatic Role Assignments

When users login, they do automatically get a role assigned. The automatic role assignment can be configured in different ways. The PROXY_ROLE_ASSIGNMENT_DRIVER environment variable (or the driver setting in the role_assignment section of the configuration file select which mechanism to use for the automatic role assignment.

When set to default, all users which do not have a role assigned at the time for the first login will get the role 'user' assigned. (This is also the default behavior if PROXY_ROLE_ASSIGNMENT_DRIVER is unset.

When PROXY_ROLE_ASSIGNMENT_DRIVER is set to oidc the role assignment for a user will happen based on the values of an OpenID Connect Claim of that user. The name of the OpenID Connect Claim to be used for the role assignment can be configured via the PROXY_ROLE_ASSIGNMENT_OIDC_CLAIM environment variable. It is also possible to define a mapping of claim values to role names defined in Infinite Scale via a yaml configuration. See the following proxy.yaml snippet for an example.

role_assignment:
    driver: oidc
    oidc_role_mapper:
        role_claim: ocisRoles
        role_mapping:
            - role_name: admin
              claim_value: myAdminRole
            - role_name: spaceadmin
              claim_value: mySpaceAdminRole
            - role_name: user
              claim_value: myUserRole
            - role_name: guest
              claim_value: myGuestRole

This would assign the role admin to users with the value myAdminRole in the claim ocisRoles. The role user to users with the values myUserRole in the claims ocisRoles and so on.

Claim values that are not mapped to a specific ownCloud Infinite Scale role will be ignored.

Note: An ownCloud Infinite Scale user can only have a single role assigned. If the configured role_mapping and a user's claim values result in multiple possible roles for a user, the order in which the role mappings are defined in the configuration is important. The first role in the role_mappings where the claim_value matches a value from the user's roles claim will be assigned to the user. So if e.g. a user's ocisRoles claim has the values myUserRole and mySpaceAdminRole that user will get the ocis role spaceadmin assigned (because spaceadmin appears before user in the above sample configuration).

If a user's claim values don't match any of the configured role mappings an error will be logged and the user will not be able to login.

The default role_claim (or PROXY_ROLE_ASSIGNMENT_OIDC_CLAIM) is roles. The default role_mapping is:

- role_name: admin
  claim_value: ocisAdmin
- role_name: spaceadmin
  claim_value: ocisSpaceAdmin
- role_name: user
  claim_value: ocisUser
- role_name: guest
  claim_value: ocisGuest

Recommendations for Production Deployments

In a production deployment, you want to have basic authentication (PROXY_ENABLE_BASIC_AUTH) disabled which is the default state. You also want to setup a firewall to only allow requests to the proxy service or the reverse proxy if you have one. Requests to the other services should be blocked by the firewall.

Content Security Policy

For Infinite Scale, external resources like an IDP (e.g. Keycloak) or when using web office documents or web apps, require defining a CSP. If not defined, the referenced services will not work.

To create a Content Security Policy (CSP), you need to create a yaml file containing the CSP definitions. To activate the settings, reference the file as value in the PROXY_CSP_CONFIG_FILE_LOCATION environment variable. For each change, a restart of the Infinite Scale deployment or the proxy service is required.

A working example for a CSP can be found in a sub path of the config directory of the ocis_full deployment example.

See the Content Security Policy (CSP) Quick Reference Guide for a description of directives.

Caching

The proxy service can use a configured store via PROXY_OIDC_USERINFO_CACHE_STORE. Possible stores are:

  • memory: Basic in-memory store and the default.
  • redis-sentinel: Stores data in a configured Redis Sentinel cluster.
  • nats-js-kv: Stores data using key-value-store feature of nats jetstream
  • noop: Stores nothing. Useful for testing. Not recommended in production environments.

Other store types may work but are not supported currently.

Note: The service can only be scaled if not using memory store and the stores are configured identically over all instances!

Note that if you have used one of the deprecated stores, you should reconfigure to one of the supported ones as the deprecated stores will be removed in a later version.

Store specific notes:

  • When using redis-sentinel, the Redis master to use is configured via e.g. OCIS_CACHE_STORE_NODES in the form of <sentinel-host>:<sentinel-port>/<redis-master> like 10.10.0.200:26379/mymaster.
  • When using nats-js-kv it is recommended to set OCIS_CACHE_STORE_NODES to the same value as OCIS_EVENTS_ENDPOINT. That way the cache uses the same nats instance as the event bus.
  • When using the nats-js-kv store, it is possible to set OCIS_CACHE_DISABLE_PERSISTENCE to instruct nats to not persist cache data on disc.

Presigned Urls

To authenticate presigned URLs the proxy service needs to read signing keys from a store that is populated by the ocs service. Possible stores are:

  • nats-js-kv: Stores data using key-value-store feature of nats jetstream
  • redis-sentinel: Stores data in a configured Redis Sentinel cluster.
  • ocisstoreservice: Stores data in the legacy ocis store service. Requires setting PROXY_PRESIGNEDURL_SIGNING_KEYS_STORE_NODES to com.owncloud.api.store.

The memory store cannot be used as it does not share the memory from the ocs service signing key memory store, even in a single process.

Make sure to configure the same store in the ocs service.

Store specific notes:

  • When using redis-sentinel, the Redis master to use is configured via e.g. OCIS_CACHE_STORE_NODES in the form of <sentinel-host>:<sentinel-port>/<redis-master> like 10.10.0.200:26379/mymaster.
  • When using nats-js-kv it is recommended to set OCS_PRESIGNEDURL_SIGNING_KEYS_STORE_NODES to the same value as PROXY_PRESIGNEDURL_SIGNING_KEYS_STORE_NODES. That way the ocs uses the same nats instance as the proxy service.
  • When using the nats-js-kv store, it is possible to set PROXY_PRESIGNEDURL_SIGNING_KEYS_STORE_DISABLE_PERSISTENCE to instruct nats to not persist signing key data on disc.
  • When using ocisstoreservice the PROXY_PRESIGNEDURL_SIGNING_KEYS_STORE_NODES must be set to the service name com.owncloud.api.store. It does not support TTL and stores the presigning keys indefinitely. Also, the store service needs to be started.

Special Settings

When using the ocis IDP service instead of an external IDP:

  • Use the environment variable OCIS_URL to define how ocis can be accessed, mandatory use https as protocol for the URL.
  • If no reverse proxy is set up, the PROXY_TLS environment variable must be set to true because the embedded libreConnect shipped with the IDP service has a hard check if the connection is on TLS and uses the HTTPS protocol. If this mismatches, an error will be logged and no connection from the client can be established.
  • PROXY_TLS can be set to false if a reverse proxy is used and the https connection is terminated at the reverse proxy. When setting to false, the communication between the reverse proxy and ocis is not secured. If set to true, you must provide certificates.

Metrics

The proxy service in ocis has the ability to expose metrics in the prometheus format. The metrics are exposed on the /metrics endpoint. There are two ways to run the ocis proxy service which has an impact on the number of metrics exposed.

1) Single Process Mode

In the single process mode, all ocis services are running inside a single process. This is the default mode when using the ocis server command to start the services. In this mode, the proxy service exposes metrics about the proxy service itself and about the ocis services it is proxying. This is due to the nature of the prometheus registry which is a singleton. The metrics exposed by the proxy service itself are prefixed with ocis_proxy_ and the metrics exposed by other ocis services are prefixed with ocis_<service-name>_.

2) Standalone Mode

In this mode, the proxy service only exposes its own metrics. The metrics of the other ocis services are exposed on their own metrics endpoints.

Available Metrics

The following metrics are exposed by the proxy service:

Metric Name Description Labels
ocis_proxy_requests_total Counter metric which reports the total number of HTTP requests. method: HTTP method of the request
ocis_proxy_errors_total Counter metric which reports the total number of HTTP requests which have failed. That counts all response codes >= 500 method: HTTP method of the request
ocis_proxy_duration_seconds Histogram of the time (in seconds) each request took. A histogram metric uses buckets to count the number of events that fall into each bucket. method: HTTP method of the request
ocis_proxy_build_info{version} A metric with a constant 1 value labeled by version, exposing the version of the ocis proxy service. version: Build version of the proxy

Prometheus Configuration

The following is an example prometheus configuration for the single process mode. It assumes that the proxy debug address is configured to bind on all interfaces PROXY_DEBUG_ADDR=0.0.0.0:9205 and that the proxy is available via the ocis service name (typically in docker-compose). The prometheus service detects the /metrics endpoint automatically and scrapes it every 15 seconds.

global:
  scrape_interval: 15s
scrape_configs:
  - job_name: ocis_proxy
    static_configs:
    - targets: ["ocis:9205"]