Files
zitadel/internal/command/instance_member.go
T
3ff4fbc7e4 fix(eventstore): add owners to unique constraints (#12740)
<!-- CURSOR_AGENT_PR_BODY_BEGIN -->
# Which Problems Are Solved

Bulk unique-constraint deletes (org, IDP with many links, user, project)
reconstruct every uniqueness key in Go and emit N `DELETE`s. That can
OOM on large tenants.

A one-shot owner backfill is also not enough: during a rolling deploy,
older binaries still insert `owners = '{}'`, so a later owner-only
delete would miss those rows.

Owner-only org/user/grant delete only matches rows whose `owners` bag
contains that tag. Several grant and member types stamped the aggregate
resource owner, which is not always the user’s home org or the granted
org. Username-scope rewrites also dropped `user:` tags, so owner-only
user delete could miss rewritten username rows.

Command must not switch to owner-only delete on the first 79 of a
version, because leftover `'{}'` rows from rolling-update overlap are
only backfilled on a later 79. A failed 79 must not turn owner-delete
on, and backfill joins must not stamp the wrong org (unverified domains,
org-scoped usernames colliding with instance-scoped ones).

# How the Problems Are Solved

Uniqueness stays the existing primary key `(instance_id, unique_type,
unique_field)`. `owners` is only a lifecycle bag of tags (`org:<id>`,
`user:<id>`, `idp:<id>`, `project:<id>`, `grant:<id>`).

- Add `owners TEXT[] NOT NULL DEFAULT '{}'` on
`eventstore.unique_constraints` (init SQL for greenfield, setup 78 for
existing clusters).
- Create `unique_constraints_owners_gin` with `CREATE INDEX
CONCURRENTLY` **outside a transaction** (same pattern as step 76).
Built-in `GIN (owners)` only; no `btree_gin`.
- Stamp `WithOwners` on **add** events. Field-level removes ignore
owners. `RemoveByOwner` stores the tag in `Owners` and both v1/v3 and v2
pushers run `DELETE … WHERE instance_id = $1 AND owners @>
ARRAY[$tag]::text[]`.
- Setup 79 backfills `owners` from projections via numbered
`cmd/setup/79/*.sql` files. Table names are interpolated from
`internal/query/projection` constants. Matches use `unique_field IN (x,
lower(x))` so the primary key can be used. Missing projection tables
(`42P01`) are skipped.
- Org domains are tagged only when `is_verified`. Instance-scoped
usernames join only users that are not org-scoped (organization settings
or domain policy `user_login_must_be_domain`). Every username row still
gets `org:` + `user:` of that user.
- 79 is a **repeatable** migration. `Check` re-runs when
`lastRun["version"]` changed or when setup-only `ForceFinalize` is set
and `lastRun["finalized"]` is still false. `Finalized` is written only
after Execute succeeds.
- Failed setup events keep the last successful `lastRun`. Command reads
`finalized` only from `system.migration.repeatable.done`.
- Org / user / project remove skip N-key event replay **when that
`finalized` bit is true**. Otherwise they keep today’s N-key lists
**and** still emit `RemoveByOwner`. The “ready” result is cached in
memory once true.

Setup writes `finalized` on the 79 repeatable.done payload:

- First 79 of a version: backfill only, `finalized=false` (unless
`ForceFinalize` is already set).
- Later 79: version change or
`BackfillUniqueConstraintOwners.ForceFinalize` /
`ZITADEL_BACKFILLUNIQUECONSTRAINTOWNERS_FORCEFINALIZE` after the rolling
update is done. That run backfills leftovers and writes
`finalized=true`.

Member and grant owner bags now include every org that should delete the
row:

- Instance, org, and project members also stamp
`org:<user_resource_owner>`.
- User grants stamp grant, user, project, and granted orgs.
- Project grant members stamp the project org, user org, and granted
org.
- Project grants stamp both orgs plus `grant:<id>`. Grant remove emits
`RemoveByOwner(grant)`.
- Username-scope rewrites keep `user:<id>` as well as `org:`.
- User/grant orgs are carried as unexported add-event fields (not
persisted JSON). `WithOwners` skips empty and duplicate tags.

# Additional Changes

- Table-driven owner-tag tests, including project grant (two `org:` tags
plus `grant:`), cross-org members, user grants, and username-scope
rewrite `user:` tags.
- Table-driven 79 `Check` + `finalizeAfterSuccess` tests for version
change, sequential version bumps, ForceFinalize restart, and
already-finalized skip.
- Command tests: a failed 79 event with `finalized=true` does not enable
owner-delete.
- Integration fixtures (verified vs unverified org domain;
instance-scoped `bob` vs org-scoped `bob`).
- Instance-only and global unique types stay untagged (`'{}'`). Instance
wipe still uses `UniqueConstraintInstanceRemove`.
- Pusher owner cases live in `eventstore_pusher_owners_test.go` so
`eventstore_pusher_test.go` stays under 1k lines.

# Additional Context

A later cleanup can delete the N-key replay helpers once a release that
ran 79 a second time (or an operator who set `ForceFinalize`) is in
production. That is not required for correctness.

<!-- CURSOR_AGENT_PR_BODY_END -->

<div><a
href="https://cursor.com/agents/bc-44f9bc9b-9a93-4cd9-8dd1-663a7b9aad2a?cursor_ref=pr_footer&cursor_cta=open_in_web"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://cursor.com/assets/images/open-in-web-dark.png"><source
media="(prefers-color-scheme: light)"
srcset="https://cursor.com/assets/images/open-in-web-light.png"><img
alt="Open in Web" width="114" height="28"
src="https://cursor.com/assets/images/open-in-web-dark.png"></picture></a>&nbsp;<a
href="https://cursor.com/background-agent?bcId=bc-44f9bc9b-9a93-4cd9-8dd1-663a7b9aad2a&cursor_ref=pr_footer&cursor_cta=open_in_cursor"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://cursor.com/assets/images/open-in-cursor-dark.png"><source
media="(prefers-color-scheme: light)"
srcset="https://cursor.com/assets/images/open-in-cursor-light.png"><img
alt="Open in Cursor" width="131" height="28"
src="https://cursor.com/assets/images/open-in-cursor-dark.png"></picture></a>&nbsp;</div>

---------

Co-authored-by: Cursor Agent <[email protected]>
Co-authored-by: Silvan <[email protected]>
2026-09-18 15:08:42 +02:00

287 lines
9.7 KiB
Go

package command
import (
"context"
"slices"
"github.com/zitadel/zitadel/internal/api/authz"
"github.com/zitadel/zitadel/internal/command/preparation"
"github.com/zitadel/zitadel/internal/domain"
"github.com/zitadel/zitadel/internal/eventstore"
"github.com/zitadel/zitadel/internal/repository/instance"
"github.com/zitadel/zitadel/internal/telemetry/tracing"
"github.com/zitadel/zitadel/internal/zerrors"
)
func (c *Commands) AddInstanceMemberCommand(a *instance.Aggregate, userID string, roles ...string) preparation.Validation {
return func() (preparation.CreateCommands, error) {
if userID == "" {
return nil, zerrors.ThrowInvalidArgument(nil, "INSTA-SDSfs", "Errors.Invalid.Argument")
}
if len(domain.CheckForInvalidRoles(roles, domain.IAMRolePrefix, c.zitadelRoles)) > 0 {
return nil, zerrors.ThrowInvalidArgument(nil, "INSTANCE-4m0fS", "Errors.Instance.MemberInvalid")
}
return func(ctx context.Context, filter preparation.FilterToQueryReducer) ([]eventstore.Command, error) {
userResourceOwner, exists, err := existingUser(ctx, filter, userID, "", false)
if err != nil || !exists {
return nil, zerrors.ThrowPreconditionFailed(err, "INSTA-GSXOn", "Errors.User.NotFound")
}
if isMember, err := IsInstanceMember(ctx, filter, a.ID, userID); err != nil || isMember {
return nil, zerrors.ThrowAlreadyExists(err, "INSTA-pFDwe", "Errors.Instance.Member.AlreadyExists")
}
return []eventstore.Command{instance.NewMemberAddedEvent(ctx, &a.Aggregate, userID, roles...).WithUserResourceOwner(userResourceOwner)}, nil
},
nil
}
}
func IsInstanceMember(ctx context.Context, filter preparation.FilterToQueryReducer, instanceID, userID string) (isMember bool, err error) {
events, err := filter(ctx, eventstore.NewSearchQueryBuilder(eventstore.ColumnsEvent).
OrderAsc().
AddQuery().
AggregateIDs(instanceID).
AggregateTypes(instance.AggregateType).
EventTypes(
instance.MemberAddedEventType,
instance.MemberRemovedEventType,
instance.MemberCascadeRemovedEventType,
).Builder())
if err != nil {
return false, err
}
for _, event := range events {
switch e := event.(type) {
case *instance.MemberAddedEvent:
if e.UserID == userID {
isMember = true
}
case *instance.MemberRemovedEvent:
if e.UserID == userID {
isMember = false
}
case *instance.MemberCascadeRemovedEvent:
if e.UserID == userID {
isMember = false
}
}
}
return isMember, nil
}
type AddInstanceMember struct {
InstanceID string
UserID string
Roles []string
}
// AddInstanceMember adds a user as a member of the instance with the given roles.
func (c *Commands) AddInstanceMember(ctx context.Context, member *AddInstanceMember) (*domain.ObjectDetails, error) {
if err := c.checkPermissionUpdateInstanceMember(ctx, member.InstanceID); err != nil {
return nil, err
}
return c.addInstanceMember(ctx, member)
}
// EnsureInstanceMemberRolesFromLogin makes sure the user holds the given instance member roles
// during the login v1 ZITADEL-IdP flow.
// It adds the membership if the user is not a member yet and otherwise adds the missing roles
// to the existing membership. Roles the user already holds are never removed.
//
// It intentionally bypasses the standard instance-member permission check.
// Authorization is established by the validated `urn:zitadel:iam:org:project:roles` token claim
// and the configured InstanceRolesInfo for the Zitadel provider.
// Do not use it for any user-initiated API path.
func (c *Commands) EnsureInstanceMemberRolesFromLogin(ctx context.Context, member *AddInstanceMember) (*domain.ObjectDetails, error) {
if member.InstanceID == "" || member.UserID == "" {
return nil, zerrors.ThrowInvalidArgument(nil, "INSTA-Ee7ai", "Errors.Instance.MemberInvalid")
}
granted := configuredRolesOnly(member.Roles, c.zitadelRoles)
if len(granted) == 0 {
return nil, zerrors.ThrowInvalidArgument(nil, "INSTA-oo0Ai", "Errors.Instance.MemberInvalid")
}
existingMember, err := c.instanceMemberWriteModelByID(ctx, member.InstanceID, member.UserID)
if err != nil {
return nil, err
}
if !existingMember.State.Exists() {
return c.addInstanceMember(ctx, &AddInstanceMember{
InstanceID: member.InstanceID,
UserID: member.UserID,
Roles: granted,
})
}
// the roles are merged so that roles granted elsewhere are retained.
roles := missingRolesAdded(existingMember.Roles, granted)
if len(roles) == len(existingMember.Roles) {
return writeModelToObjectDetails(&existingMember.WriteModel), nil
}
pushedEvents, err := c.eventstore.Push(ctx,
instance.NewMemberChangedEvent(ctx,
InstanceAggregateFromWriteModel(&existingMember.WriteModel),
member.UserID,
roles...,
),
)
if err != nil {
return nil, err
}
if err = AppendAndReduce(existingMember, pushedEvents...); err != nil {
return nil, err
}
return writeModelToObjectDetails(&existingMember.WriteModel), nil
}
// configuredRolesOnly returns the subset of roles that are configured as instance roles.
// The roles of a ZITADEL provider's claim originate from an external
// instance, whose role set may differ, so unknown roles are dropped.
func configuredRolesOnly(roles []string, zitadelRoles []authz.RoleMapping) []string {
invalid := domain.CheckForInvalidRoles(roles, domain.IAMRolePrefix, zitadelRoles)
return slices.DeleteFunc(slices.Clone(roles), func(role string) bool {
return slices.Contains(invalid, role)
})
}
// missingRolesAdded returns the existing roles with new roles to be added that are not already
// present, preserving the order of the existing roles.
func missingRolesAdded(existing, add []string) []string {
roles := slices.Clone(existing)
for _, role := range add {
if !slices.Contains(roles, role) {
roles = append(roles, role)
}
}
return roles
}
// addInstanceMember performs the write without any permission check.
func (c *Commands) addInstanceMember(ctx context.Context, member *AddInstanceMember) (*domain.ObjectDetails, error) {
instanceAgg := instance.NewAggregate(member.InstanceID)
//nolint:staticcheck
cmds, err := preparation.PrepareCommands(ctx, c.eventstore.Filter, c.AddInstanceMemberCommand(instanceAgg, member.UserID, member.Roles...))
if err != nil {
return nil, err
}
events, err := c.eventstore.Push(ctx, cmds...)
if err != nil {
return nil, err
}
addedMember := NewInstanceMemberWriteModel(member.InstanceID, member.UserID)
err = AppendAndReduce(addedMember, events...)
if err != nil {
return nil, err
}
return writeModelToObjectDetails(&addedMember.WriteModel), nil
}
type ChangeInstanceMember struct {
InstanceID string
UserID string
Roles []string
}
func (i *ChangeInstanceMember) IsValid(zitadelRoles []authz.RoleMapping) error {
if i.InstanceID == "" || i.UserID == "" || len(i.Roles) == 0 {
return zerrors.ThrowInvalidArgument(nil, "INSTANCE-LiaZi", "Errors.Instance.MemberInvalid")
}
if len(domain.CheckForInvalidRoles(i.Roles, domain.IAMRolePrefix, zitadelRoles)) > 0 {
return zerrors.ThrowInvalidArgument(nil, "INSTANCE-3m9fs", "Errors.Instance.MemberInvalid")
}
return nil
}
// ChangeInstanceMember updates an existing member
func (c *Commands) ChangeInstanceMember(ctx context.Context, member *ChangeInstanceMember) (*domain.ObjectDetails, error) {
if err := member.IsValid(c.zitadelRoles); err != nil {
return nil, err
}
existingMember, err := c.instanceMemberWriteModelByID(ctx, member.InstanceID, member.UserID)
if err != nil {
return nil, err
}
if !existingMember.State.Exists() {
return nil, zerrors.ThrowNotFound(nil, "INSTANCE-D8JxR", "Errors.NotFound")
}
if err := c.checkPermissionUpdateInstanceMember(ctx, existingMember.AggregateID); err != nil {
return nil, err
}
if slices.Compare(existingMember.Roles, member.Roles) == 0 {
return writeModelToObjectDetails(&existingMember.WriteModel), nil
}
pushedEvents, err := c.eventstore.Push(ctx,
instance.NewMemberChangedEvent(ctx,
InstanceAggregateFromWriteModel(&existingMember.WriteModel),
member.UserID,
member.Roles...,
),
)
if err != nil {
return nil, err
}
err = AppendAndReduce(existingMember, pushedEvents...)
if err != nil {
return nil, err
}
return writeModelToObjectDetails(&existingMember.WriteModel), nil
}
func (c *Commands) RemoveInstanceMember(ctx context.Context, instanceID, userID string) (*domain.ObjectDetails, error) {
if userID == "" {
return nil, zerrors.ThrowInvalidArgument(nil, "INSTANCE-LiaZi", "Errors.IDMissing")
}
existingMember, err := c.instanceMemberWriteModelByID(ctx, instanceID, userID)
if err != nil {
return nil, err
}
if err := c.checkPermissionDeleteInstanceMember(ctx, instanceID); err != nil {
return nil, err
}
if !existingMember.State.Exists() {
return writeModelToObjectDetails(&existingMember.WriteModel), nil
}
pushedEvents, err := c.eventstore.Push(ctx,
c.removeInstanceMember(ctx,
InstanceAggregateFromWriteModel(&existingMember.WriteModel),
userID,
false,
),
)
if err != nil {
return nil, err
}
err = AppendAndReduce(existingMember, pushedEvents...)
if err != nil {
return nil, err
}
return writeModelToObjectDetails(&existingMember.WriteModel), nil
}
func (c *Commands) removeInstanceMember(ctx context.Context, instanceAgg *eventstore.Aggregate, userID string, cascade bool) eventstore.Command {
if cascade {
return instance.NewMemberCascadeRemovedEvent(
ctx,
instanceAgg,
userID)
} else {
return instance.NewMemberRemovedEvent(ctx, instanceAgg, userID)
}
}
func (c *Commands) instanceMemberWriteModelByID(ctx context.Context, instanceID, userID string) (member *InstanceMemberWriteModel, err error) {
ctx, span := tracing.NewSpan(ctx)
defer func() { span.EndWithError(err) }()
writeModel := NewInstanceMemberWriteModel(instanceID, userID)
err = c.eventstore.FilterToQueryReducer(ctx, writeModel)
if err != nil {
return nil, err
}
return writeModel, nil
}