Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

chore(deps): update dependency @apollo/gateway to v2.8.5 [security] #3310

Merged
merged 1 commit into from
Oct 6, 2024
Merged

chore(deps): update dependency @apollo/gateway to v2.8.5 [security] #3310

merged 1 commit into from
Oct 6, 2024

Conversation

renovate[bot]
Copy link
Contributor

@renovate renovate bot commented Aug 27, 2024 •
edited
Loading

This PR contains the following updates:

Package Change Age Adoption Passing Confidence
@apollo/gateway (source) 2.3.3 -> 2.8.5 age adoption passing confidence

GitHub Vulnerability Alerts

CVE-2024-43414

Impact

Instances of @​apollo/query-planner >=2.0.0 and <2.8.5 are impacted by a denial-of-service vulnerability. @​apollo/gateway versions >=2.0.0 and < 2.8.5 and Apollo Router <1.52.1 are also impacted through their use of @​apollo/query-planner.

If @​apollo/query-planner is asked to plan a sufficiently complex query, it may loop infinitely and never complete. This results in unbounded memory consumption and either a crash or out-of-memory (OOM) termination.

This issue can be triggered if you have at least one non-@key field that can be resolved by multiple subgraphs. To identify these shared fields, the schema for each subgraph must be reviewed. The mechanism to identify shared fields varies based on the version of Federation your subgraphs are using.

You can check if your subgraphs are using Federation 1 or Federation 2 by reviewing their schemas. Federation 2 subgraph schemas will contain a @link directive referencing the version of Federation being used while Federation 1 subgraphs will not. For example, in a Federation 2 subgraph, you will find a line like @link(url: "https://specs.apollo.dev/federation/v2.0"). If a similar @link directive is not present in your subgraph schema, it is using Federation 1. Note that a supergraph can contain a mix of Federation 1 and Federation 2 subgraphs.

To review Federation 1 subgraphs for impact:

In Federation 1 subgraphs, fields are implicitly shareable across subgraphs. To review for impact, you will need to review for cases where multiple subgraphs can resolve the same field. For example:

# Subgraph 1
type Query {
  field: Int
}

# Subgraph 2
type Query {
  field: Int
}

To review Federation 2 subgraphs for impact:

In Federation 2 subgraphs, fields must be explicitly defined as shareable across subgraphs. This is done via the @shareable directive. For example:

# Subgraph 1
@&#8203;link(url: "https://specs.apollo.dev/federation/v2.0")
type Query {
  field: Int @&#8203;shareable
}

# Subgraph 2
@&#8203;link(url: "https://specs.apollo.dev/federation/v2.0")
type Query {
  field: Int @&#8203;shareable
}

Impact Detail

This issue results from the Apollo query planner attempting to use a Number exceeding Javascript’s Number.MAX_VALUE in some cases. In Javascript, Number.MAX_VALUE is (2^1024 - 2^971).

When the query planner receives an inbound graphql request, it breaks the query into pieces and for each piece, generates a list of potential execution steps to solve the piece. These candidates represent the steps that the query planner will take to satisfy the pieces of the larger query. As part of normal operations, the query planner requires and calculates the number of possible query plans for the total query. That is, it needs the product of the number of query plan candidates for each piece of the query. Under normal circumstances, after generating all query plan candidates and calculating the number of all permutations, the query planner moves on to stack rank candidates and prune less-than-optimal options.

In particularly complex queries, especially those where fields can be solved through multiple subgraphs, this can cause the number of all query plan permutations to balloon. In worst-case scenarios, this can end up being a number larger than Number.MAX_VALUE. In Javascript, if Number.MAX_VALUE is exceeded, Javascript represents the value as “infinity”. If the count of candidates is evaluated as infinity, the component of the query planner responsible for pruning less-than-optimal query plans does not actually prune candidates, causing the query planner to evaluate many orders of magnitude more query plan candidates than necessary.

A given graph’s exposure to this issue varies based on its complexity. Consider the following Federation 2 subgraphs:

# Subgraph 1
type Query {
  field: Int @&#8203;shareable
}

# Subgraph 2
type Query {
  field: Int @&#8203;shareable
}

The query planner can solve requests for Query.field in one of two ways - either by querying subgraph 1 or subgraph 2.

The following query with 1024 aliased fields would trigger this issue because 2^1024 > Number.MAX_VALUE:

query {
  field_1: field
  field_2: field
  # ...
  field_1023: field
  field_1024: field
}

However, in a graph that provided 5 options to solve a given field, the bug could be encountered in a query that aliased the field approximately 440 times.

Patches

@​apollo/query-planner 2.8.5
@​apollo/gateway 2.8.5
Apollo Router 1.52.1

Workarounds

This issue can be avoided by ensuring there are no fields resolvable from multiple subgraphs. If all subgraphs are using Federation 2, you can confirm that you are not impacted by ensuring that none of your subgraph schemas use the @shareable directive. If you are using Federation 1 subgraphs, you will need to validate that there are no fields resolvable by multiple subgraphs.

Note that a supergraph can contain a mix of Federation 1 and Federation 2 subgraphs.

If you do have fields resolvable by multiple subgraphs, changing this behavior in response to this issue may be risky to the operation of your supergraph. We recommend that you update to a patched version of either Apollo Router or Apollo Gateway.

Apollo customers with an enterprise entitlement using the Apollo Router can also mitigate much of the risk from this issue by implementing Apollo’s Persisted Queries (PQ) feature. With PQ enabled, the Apollo Router will only execute safelisted queries. While customers would need to ensure that queries that induce this issue are not added to the safelist, PQs would mitigate the risk of clients submitting ad hoc queries that exploit this issue.

References

Additional information on Query Plans

Release Notes

apollographql/federation (@​apollo/gateway)

v2.8.5

Compare Source

v2.8.4

Compare Source

Patch Changes

v2.8.3

Compare Source

Patch Changes

v2.8.2

Compare Source

Patch Changes

v2.8.1

Compare Source

Patch Changes

v2.8.0

Compare Source

Minor Changes

  • Implement new directives to allow getting and setting context. This allows resolvers to reference and access data referenced by entities that exist in the GraphPath that was used to access the field. The following example demonstrates the ability to access the prop field within the Child resolver. (#​2988)

    type Query {
  p: Parent!
}
type Parent @&#8203;key(fields: "id") @&#8203;context(name: "context") {
  id: ID!
  child: Child!
  prop: String!
}
type Child @&#8203;key(fields: "id") {
  id: ID!
  b: String!
  field(a: String @&#8203;fromContext(field: "$context { prop }")): Int!
}
Patch Changes

v2.7.8

Compare Source

Patch Changes

v2.7.7

Compare Source

Patch Changes

v2.7.6

Compare Source

Patch Changes

v2.7.5

Compare Source

Patch Changes

v2.7.4

Compare Source

Patch Changes

v2.7.3

Compare Source

Patch Changes

v2.7.2

Compare Source

Patch Changes

v2.7.1

Compare Source

Patch Changes

v2.7.0

Compare Source

Minor Changes

  • Implement progressive @override functionality (#​2911)

    The progressive @override feature brings a new argument to the @override directive: label: String. When a label is added to an @override application, the override becomes conditional, depending on parameters provided to the query planner (a set of which labels should be overridden). Note that this feature will be supported in router for enterprise users only.

    Out-of-the-box, the router will support a percentage-based use case for progressive @override. For example:

    type Query {
  hello: String @&#8203;override(from: "original", label: "percent(5)")
}

    The above example will override the root hello field from the "original" subgraph 5% of the time.

    More complex use cases will be supported by the router via the use of coprocessors/rhai to resolve arbitrary labels to true/false values (i.e. via a feature flag service).

Patch Changes

v2.6.3

Compare Source

Patch Changes

v2.6.2

Compare Source

Patch Changes

v2.6.1

Compare Source

Patch Changes

v2.6.0

Compare Source

Minor Changes

  • Add more information to OpenTelemetry spans. (#​2700)

    Rename operationName to graphql.operation.name and add a
    graphql.operation.type attribute, in conformance with the OpenTelemetry
    Semantic Conventions for GraphQL. The operationName attribute is now
    deprecated, but it is still emitted alongside graphql.operation.name.

    Add a graphql.document span attribute to the gateway.request span,
    containing the entire GraphQL source sent in the request. This feature
    is disable by default.

    When one or more GraphQL or internal errors occur, report them in the
    OpenTelemetry span in which they took place, as an exception event. This
    feature is disabled by default.

    To enable the graphql.document span attribute and the exception event
    reporting, add the following entries to your ApolloGateway instance
    configuration:

    const gateway = new ApolloGateway({
  // ...
  telemetry: {
    // Set to `true` to include the `graphql.document` attribute
    includeDocument: true,
    // Set to `true` to report all exception events, or set to a number
    // to report at most that number of exception events per span
    reportExceptions: true,
    // or: reportExceptions: 1
  },
});

  • Update license field in package.json to use Elastic-2.0 SPDX identifier (#​2741)

  • Introduce the new @policy scope for composition (#​2818)

    Note that this directive will only be fully supported by the Apollo Router as a GraphOS Enterprise feature at runtime. Also note that composition of valid @policy directive applications will succeed, but the resulting supergraph will not be executable by the Gateway or an Apollo Router which doesn't have the GraphOS Enterprise entitlement.

    Users may now compose @policy applications from their subgraphs into a supergraph.

    The directive is defined as follows:

    scalar federation__Policy

directive @&#8203;policy(
  policies: [[federation__Policy!]!]!
) on FIELD_DEFINITION | OBJECT | INTERFACE | SCALAR | ENUM

    The Policy scalar is effectively a String, similar to the FieldSet type.

    In order to compose your @policy usages, you must update your subgraph's federation spec version to v2.6 and add the @policy import to your existing imports like so:

    @&#8203;link(url: "https://specs.apollo.dev/federation/v2.6", import: [..., "@&#8203;policy"])

  • Add graphql.operation.name attribute on gateway.plan span (#​2807)

Patch Changes

v2.5.7

Compare Source

Patch Changes

v2.5.6

Compare Source

Patch Changes

v2.5.5

Compare Source

Patch Changes

  • Fix specific case for requesting __typename on interface entity type (#​2775)

    In certain cases, when resolving a __typename on an interface entity (due to it actual being requested in the operation), that fetch group could previously be trimmed / treated as useless. At a glance, it appears to be a redundant step, i.e.:

    { ... on Product { __typename id }} => { ... on Product { __typename} }

    It's actually necessary to preserve this in the case that we're coming from an interface object to an (entity) interface so that we can resolve the concrete __typename correctly.

  • Don't preserve useless fetches which downgrade __typename from a concrete type back to its interface type. (#​2778)

    In certain cases, the query planner was preserving some fetches which were "useless" that would rewrite __typename from its already-resolved concrete type back to its interface type. This could result in (at least) requested fields being "filtered" from the final result due to the interface's __typename in the data where the concrete type's __typename was expected.

    Specifically, the solution was compute the path between newly created groups and their parents when we know that it's trivial ([]). Further along in the planning process, this allows to actually remove the known-useless group.

  • Propagate type information when renaming entity fields (#​2776)

    Aliased entity fields might have been incorrectly overwritten if multiple fields/aliases shared the same name. Query planner automatically renames conflicting fields to ensure we can always generate a valid GraphQL query. The underlying issue was that this key rewriting logic was assuming the same type of an object. In case of entity queries asking for those aliased fields, we ended up always attempting to apply field renaming logic regardless, whether or not a given entity was of the correct type. This fix ensures that the query planner logic correctly accounts for the object type when applying field renaming logic.

  • Updated dependencies [66d7e4ce, a37bbbf6]:

v2.5.4

Compare Source

Patch Changes

  • Adds header to change the format of exposed query plans, and allows formatting it as json. (#​2724)

    When the gateway is configured to allow it, adding the Apollo-Query-Plan-Experimental header to a request already allowed a "prettified" text version of the query plan used for the query is returned in the response extension. This changes adds support for a new (optional) accompanying header, Apollo-Query-Plan-Experimental-Format, which can be set to the value "internal" to have the query plan returned as a json object (that correspond to the internal representation of that query plan) instead of the text version otherwise sent. Note that if that new header is not provided, then the query plan continues to be send in the previous prettified text version.

  • Fix some potentially incorrect query plans with @requires when some dependencies are involved. (#​2726)

    In some rare case of @requires, an over-eager optimisation was incorrectly considering that
    a dependency between 2 subgraph fetches was unnecessary, leading to doing 2 subgraphs queries
    in parallel when those should be done sequentially (because the 2nd query rely on results
    from the 1st one). This effectively resulted in the required fields not being provided (the
    consequence of which depends a bit on the resolver detail, but if the resolver expected
    the required fields to be populated (as they should), then this could typically result
    in a message of the form GraphQLError: Cannot read properties of null).

  • Updated dependencies [203b0a44]:

v2.5.3

Compare Source

Patch Changes

v2.5.2

Compare Source

Patch Changes

  • Remove extraneous call to span.setStatus() on a span which has already ended. (#​2697)

    In cases where a subgraph responded with an error, we would sometimes try to set
    the status of a span which had already ended. This resulted in a warning log to
    the console (but no effect otherwise). This warning should no longer happen.

  • Fix fallbackPollIntervalInMs behavior. (#​2709)

    The fallbackPollIntervalInMs serves 2 purposes:

    • it allows users to provide an Uplink poll interval if Uplink doesn't provide one
    • it allows users to use a longer poll interval that what's prescribed by Uplink

    The second bullet is how the configuration option is documented, but not how it was previously implemented. This change corrects the behavior to respect this configuration if it's provided AND is longer than the Uplink interval.

  • Updated dependencies [35179f08]:

v2.5.1

Compare Source

Patch Changes

  • Reapply #​2639: (#​2687)

    Try reusing named fragments in subgraph fetches even if those fragment only apply partially to the subgraph. Before this change, only named fragments that were applying entirely to a subgraph were tried, leading to less reuse that expected. Concretely, this change can sometimes allow the generation of smaller subgraph fetches.

    Additionally, resolve a bug which surfaced in the fragment optimization logic which could result in invalid/incorrect optimizations / fragment reuse.

  • Updated dependencies [b9052fdd]:

v2.5.0

Compare Source

Minor Changes

  • Do not run the full suite of graphQL validations on supergraphs and their extracted subgraphs by default in production environment. (#​2657)

    Running those validations on every updates of the schema takes a non-negligible amount of time (especially on large
    schema) and mainly only serves in catching bugs early in the supergraph handling code, and in some limited cases,
    provide slightly better messages when a corrupted supergraph is received, neither of which is worth the cost in
    production environment.

    A new validateSupergraph option is also introduced in the gateway configuration to force this behaviour.

  • Support responses from subgraphs which use the application/graphql-response+json content-type header. (#​2162)

    See graphql-over-http spec for more details:
    https://graphql.github.io/graphql-over-http/draft/#sec-application-graphql-response-json

  • Introduce the new @authenticated directive for composition (#​2644)

    Note that this directive will only be fully supported by the Apollo Router as a GraphOS Enterprise feature at runtime. Also note that composition of valid @authenticated directive applications will succeed, but the resulting supergraph will not be executable by the Gateway or an Apollo Router which doesn't have the GraphOS Enterprise entitlement.

    Users may now compose @authenticated applications from their subgraphs into a supergraph. This addition will support a future version of Apollo Router that enables authenticated access to specific types and fields via directive applications.

    The directive is defined as follows:

    directive @&#8203;authenticated on FIELD_DEFINITION | OBJECT | INTERFACE | SCALAR | ENUM

    In order to compose your @authenticated usages, you must update your subgraph's federation spec version to v2.5 and add the @authenticated import to your existing imports like so:

    @&#8203;link(url: "https://specs.apollo.dev/federation/v2.5", import: [..., "@&#8203;authenticated"])

  • Introduce the new @requiresScopes directive for composition (#​2649)

    Note that this directive will only be fully supported by the Apollo Router as a GraphOS Enterprise feature at runtime. Also note that composition of valid @requiresScopes directive applications will succeed, but the resulting supergraph will not be executable by the Gateway or an Apollo Router which doesn't have the GraphOS Enterprise entitlement.

    Users may now compose @requiresScopes applications from their subgraphs into a supergraph. This addition will support a future version of Apollo Router that enables scoped access to specific types and fields via directive applications.

    The directive is defined as follows:

    scalar federation__Scope

directive @&#8203;requiresScopes(
  scopes: [federation__Scope!]!
) on FIELD_DEFINITION | OBJECT | INTERFACE | SCALAR | ENUM

    The Scope scalar is effectively a String, similar to the FieldSet type.

    In order to compose your @requiresScopes usages, you must update your subgraph's federation spec version to v2.5 and add the @requiresScopes import to your existing imports like so:

    @&#8203;link(url: "https://specs.apollo.dev/federation/v2.5", import: [..., "@&#8203;requiresScopes"])
Patch Changes

v2.4.13

Compare Source

Patch Changes

v2.4.12

Compare Source

Patch Changes

  • Remove extraneous call to span.setStatus() on a span which has already ended. (#​2717)

    In cases where a subgraph responded with an error, we would sometimes try to set
    the status of a span which had already ended. This resulted in a warning log to
    the console (but no effect otherwise). This warning should no longer happen.

  • Fix fallbackPollIntervalInMs behavior. (#​2717)

    The fallbackPollIntervalInMs serves 2 purposes:

    • it allows users to provide an Uplink poll interval if Uplink doesn't provide one
    • it allows users to use a longer poll interval that what's prescribed by Uplink

    The second bullet is how the configuration option is documented, but not how it was previously implemented. This change corrects the behavior to respect this configuration if it's provided AND is longer than the Uplink interval.

  • Updated dependencies [693c2433]:

v2.4.11

Compare Source

Patch Changes

  • Reapply #​2639: (#​2684)

    Try reusing named fragments in subgraph fetches even if those fragment only apply partially to the subgraph. Before this change, only named fragments that were applying entirely to a subgraph were tried, leading to less reuse that expected. Concretely, this change can sometimes allow the generation of smaller subgraph fetches.

    Additionally, resolve a bug which surfaced in the fragment optimization logic which could result in invalid/incorrect optimizations / fragment reuse.

  • Updated dependencies [a740e071]:

v2.4.10

Compare Source

Patch Changes

v2.4.9

Compare Source

Patch Changes

v2.4.8

Compare Source

Patch Changes

v2.4.7

Compare Source

Patch Changes

v2.4.6

Compare Source

Patch Changes

Configuration

📅 Schedule: Branch creation - "" (UTC), Automerge - At any time (no schedule defined).

🚦 Automerge: Enabled.

♻ Rebasing: Whenever PR is behind base branch, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about this update again.

  • If you want to rebase/retry this PR, check this box

This PR was generated by Mend Renovate. View the repository job log.

@renovate renovate bot added the dependencies Pull requests that update a dependency file label Aug 27, 2024
@renovate renovate bot force-pushed the renovate/npm-apollo-gateway-vulnerability branch from faf305d to 1fa364b Compare October 6, 2024 16:09
@renovate renovate bot merged commit e665d8d into master Oct 6, 2024
1 check passed
@renovate renovate bot deleted the renovate/npm-apollo-gateway-vulnerability branch October 6, 2024 18:06
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
dependencies Pull requests that update a dependency file
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
0 participants