Merge branch 'feat/user-display-persistent-cache' into 'main'

feat(user-display): persistent DB cache + Keycloak Admin lookup (Phase 1+2)

See merge request 2-6/2-6-4/terravault/ordinis!180
This commit is contained in:
Александр Зимин
2026-05-14 11:52:20 +00:00
12 changed files with 707 additions and 47 deletions
+136
View File
@@ -0,0 +1,136 @@
# Keycloak Admin client setup for UserDisplayService Phase 2
`KeycloakAdminUserResolver` resolves a JWT `sub` against the Keycloak Admin
REST API on cache miss, so users that never hit ordinis still get a
display name in audit / reviews / changelog / events / history. Without
this, only users who personally made an authenticated request show up
with a real name (the "ток мой UUID отображается" pain).
This is a per-environment one-time setup. Disabled by default — flip on
once the Keycloak client + vault secret exist.
## Step 1: Create service-account client in Keycloak
Realm: `nstart`. Path: Clients → Create client.
| Field | Value |
|---|---|
| Client type | OpenID Connect |
| Client ID | `ordinis-admin-client` |
| Client authentication | **On** (confidential) |
| Authorization | Off |
| Standard flow | Off |
| Direct access grants | Off |
| Service accounts roles | **On** |
Save. On the created client → **Service accounts roles** tab → **Assign
role** → Filter by clients → `realm-management` → check **`view-users`**
→ Assign.
(`view-users` is the minimum scope needed for `GET /admin/realms/{realm}/users/{id}`.
Don't assign `manage-users` — read-only is enough.)
Then **Credentials** tab → copy the `Client secret` (long random string).
## Step 2: Store secret in Vault
The cluster-domain matters here — staging is `265.local`, prod is
`altum.local`. Both have separate Vault instances.
Staging (cluster.265, namespace `config`):
```bash
kubectl -n config exec -it vault-0 -- sh -c '
vault login <root-or-admin-token>
vault kv put secret/ordinis/cuod/keycloak-admin \
client_id=ordinis-admin-client \
client_secret=<paste-from-keycloak>
'
```
Prod (altum, namespace `vault`):
```bash
ssh root@192.168.100.142 'kubectl -n vault exec -it vault-0 -- sh -c "
vault login <prod-root-token>
vault kv put secret/ordinis/cuod/keycloak-admin \
client_id=ordinis-admin-client \
client_secret=<paste-from-keycloak>
"'
```
⚠️ Different secret per environment — staging and prod have separate
Keycloak clients (and certainly separate secrets).
## Step 3: Wire into ordinis-app deployment
Add to `ordinis-app` helm chart annotations (vault agent injector
template) — same pattern as `secret-postgres` / `secret-kafka` already
present in the chart:
```yaml
vault.hashicorp.com/agent-inject-secret-keycloak-admin: secret/data/ordinis/cuod/keycloak-admin
vault.hashicorp.com/agent-inject-template-keycloak-admin: |
{{- with secret "secret/data/ordinis/cuod/keycloak-admin" -}}
export ORDINIS_KEYCLOAK_ADMIN_CLIENT_ID="{{ .Data.data.client_id }}"
export ORDINIS_KEYCLOAK_ADMIN_CLIENT_SECRET="{{ .Data.data.client_secret }}"
{{- end }}
```
Add env vars to the deployment:
```yaml
env:
- name: ORDINIS_KEYCLOAK_ADMIN_ENABLED
value: "true"
- name: ORDINIS_KEYCLOAK_ADMIN_URL
value: "https://auth.nstart.space" # both envs use the same Keycloak
- name: ORDINIS_KEYCLOAK_ADMIN_REALM
value: "nstart"
# CLIENT_ID + CLIENT_SECRET arrive via the vault-injected
# /vault/secrets/keycloak-admin file sourced before java starts.
```
## Step 4: Verify
After redeploy, check the pod log on startup:
```
KeycloakAdminUserResolver active — base=https://auth.nstart.space realm=nstart client=ordinis-admin-client
```
End-to-end:
```bash
# Find a sub that's NOT in the JVM cache (e.g. a user who never logged in).
# Hit the display endpoint as an admin user:
curl -s -H "Authorization: Bearer $TOKEN" \
https://ordinis.k8s.265.nstart.cloud/api/v1/admin/users/<uuid>/display | jq
# Expect 200 with username/name/email even on cold cache, instead of 404.
```
Then check the DB — the row was persisted from the on-demand resolve:
```sql
SELECT sub, preferred_username, source, synced_at
FROM user_display_cache
WHERE sub = '<uuid>';
-- source = KEYCLOAK_ON_DEMAND
```
## Step 5: Phase 3 (later — scheduled bulk sync)
When ready to refill the table proactively (avoids the per-user
on-demand latency for fresh subs), add a Spring `@Scheduled` task that
calls `GET /admin/realms/{realm}/users` paginated and writes via
`UserDisplayService.upsertFromKeycloak`. Recommended every 6h. Not
included in this MR — depends on observed cache miss rate.
## Failure modes
- **Service-account client missing `view-users`** → 403 on lookup,
resolver throws RuntimeException, UserDisplayService logs warn,
caller falls back to short UUID. No outage, just no resolve.
- **Keycloak unreachable** → connect timeout / 502. Same fallback as
above. Cached entries (DB tier) keep working.
- **client_secret rotated in Keycloak but vault not updated** → 401 on
token request → resolver wipes cached token and surfaces
RuntimeException. Update vault secret and the next attempt succeeds.
@@ -1,6 +1,7 @@
import { useTranslation } from 'react-i18next' import { useTranslation } from 'react-i18next'
import { Badge, LoadingBlock, Modal, QueryErrorState } from '@/ui' import { Badge, LoadingBlock, Modal, QueryErrorState } from '@/ui'
import { useChangelogDiff } from '@/api/queries' import { useChangelogDiff } from '@/api/queries'
import { UserCell } from '@/lib/useUserDisplay'
import { kindBadgeVariant, kindLabel } from './kind-meta' import { kindBadgeVariant, kindLabel } from './kind-meta'
/** /**
@@ -52,9 +53,9 @@ export function ChangelogDiffModal({
{/* Header strip — kind + author */} {/* Header strip — kind + author */}
<div className="flex flex-wrap items-center gap-2"> <div className="flex flex-wrap items-center gap-2">
<Badge variant={kindBadgeVariant(data.kind)}>{kindLabel(data.kind, t)}</Badge> <Badge variant={kindBadgeVariant(data.kind)}>{kindLabel(data.kind, t)}</Badge>
<span className="text-cell text-mute"> <span className="text-cell text-mute inline-flex items-center gap-1">
{t('changelog.diff.by', { defaultValue: 'автор' })}:{' '} {t('changelog.diff.by', { defaultValue: 'автор' })}:{' '}
<span className="text-ink font-mono">{data.publishedBy.name}</span> <UserCell uuid={data.publishedBy.sub} />
</span> </span>
</div> </div>
@@ -6,6 +6,7 @@ import type { ChangelogEntry, DictionaryDetail, FlattenedRecord } from '@/api/cl
import { useChangelog, useChangelogDiff, useRecords } from '@/api/queries' import { useChangelog, useChangelogDiff, useRecords } from '@/api/queries'
import { kindBadgeVariant, kindIcon, kindLabel } from '@/components/changelog/kind-meta' import { kindBadgeVariant, kindIcon, kindLabel } from '@/components/changelog/kind-meta'
import { ChangelogDiffModal } from '@/components/changelog/ChangelogDiffModal' import { ChangelogDiffModal } from '@/components/changelog/ChangelogDiffModal'
import { UserCell } from '@/lib/useUserDisplay'
import { cn } from '@/lib/utils' import { cn } from '@/lib/utils'
/** /**
@@ -411,7 +412,9 @@ function ComparePanel({
· <span className="text-mono">{note}</span> · <span className="text-mono">{note}</span>
</span> </span>
)} )}
{author && <span className="text-mute">· {author}</span>} {author && (
<span className="text-mute inline-flex items-center gap-1">· <UserCell uuid={author} /></span>
)}
</div> </div>
</div> </div>
) )
@@ -683,7 +686,7 @@ function ChangesTabContent({
<span className="tabular-nums"> <span className="tabular-nums">
{new Date(e.publishedAt).toLocaleString()} {new Date(e.publishedAt).toLocaleString()}
</span> </span>
<span className="font-mono">{e.publishedBy.name}</span> <UserCell uuid={e.publishedBy.sub} />
<button <button
type="button" type="button"
onClick={() => onOpenDiff(e.id)} onClick={() => onOpenDiff(e.id)}
@@ -895,7 +898,7 @@ function SchemaTabContent({
<dt className="text-cap text-mute uppercase tracking-wider"> <dt className="text-cap text-mute uppercase tracking-wider">
{t('timeTravel.schema.author', { defaultValue: 'автор' })} {t('timeTravel.schema.author', { defaultValue: 'автор' })}
</dt> </dt>
<dd className="font-mono">{pickedEntry.publishedBy.name}</dd> <dd><UserCell uuid={pickedEntry.publishedBy.sub} /></dd>
<dt className="text-cap text-mute uppercase tracking-wider"> <dt className="text-cap text-mute uppercase tracking-wider">
{t('timeTravel.schema.date', { defaultValue: 'дата' })} {t('timeTravel.schema.date', { defaultValue: 'дата' })}
</dt> </dt>
@@ -56,6 +56,7 @@ import { TimeTravelModal } from '@/components/timetravel/TimeTravelModal'
import { ChangelogDiffModal } from '@/components/changelog/ChangelogDiffModal' import { ChangelogDiffModal } from '@/components/changelog/ChangelogDiffModal'
import { kindBadgeVariant, kindIcon, kindLabel, isVersioningKind } from '@/components/changelog/kind-meta' import { kindBadgeVariant, kindIcon, kindLabel, isVersioningKind } from '@/components/changelog/kind-meta'
import { CreateSchemaDraftModal } from '@/components/workflow/CreateSchemaDraftModal' import { CreateSchemaDraftModal } from '@/components/workflow/CreateSchemaDraftModal'
import { UserCell } from '@/lib/useUserDisplay'
import { nowIsoLocal } from '@/lib/dates' import { nowIsoLocal } from '@/lib/dates'
import { recordDisplayName } from '@/lib/locales' import { recordDisplayName } from '@/lib/locales'
import { SCOPE_BORDER_TOP } from '@/lib/scope-style' import { SCOPE_BORDER_TOP } from '@/lib/scope-style'
@@ -2258,8 +2259,8 @@ function EventsTabContent({ dictName }: { dictName: string }) {
{status} {status}
</span> </span>
</td> </td>
<td className="px-3 py-2 text-mono text-cell text-ink-2"> <td className="px-3 py-2 text-cell text-ink-2">
{r.userId ?? 'anonymous'} {r.userId ? <UserCell uuid={r.userId} /> : <span className="text-mono text-mute">anonymous</span>}
</td> </td>
<td className="px-3 py-2 text-cell text-ink-2 max-w-md truncate"> <td className="px-3 py-2 text-cell text-ink-2 max-w-md truncate">
{r.businessKey ? ( {r.businessKey ? (
@@ -2396,7 +2397,7 @@ function ChangelogTimelineItem({
)} )}
<div className="flex items-center gap-3 mt-1 text-cell text-mute"> <div className="flex items-center gap-3 mt-1 text-cell text-mute">
<span className="tabular-nums">{occurredAt.toLocaleString()}</span> <span className="tabular-nums">{occurredAt.toLocaleString()}</span>
<span className="font-mono">{entry.publishedBy.name}</span> <UserCell uuid={entry.publishedBy.sub} />
{entry.recordsAffected > 0 && ( {entry.recordsAffected > 0 && (
<span> <span>
{t('changelog.recordsAffected', { {t('changelog.recordsAffected', {
@@ -211,3 +211,16 @@ ordinis:
issuer: ${jwt-issuer:} issuer: ${jwt-issuer:}
require-authentication: ${ORDINIS_AUTH_REQUIRED:false} require-authentication: ${ORDINIS_AUTH_REQUIRED:false}
allow-query-scope: ${ORDINIS_AUTH_ALLOW_QUERY_SCOPE:true} allow-query-scope: ${ORDINIS_AUTH_ALLOW_QUERY_SCOPE:true}
# Keycloak Admin REST integration for UserDisplayService Phase 2 — on-demand
# sub→user lookup when JWT capture cache + DB cache both miss. See
# KeycloakAdminUserResolver.java + 0023-user-display-cache.xml migration.
# Disabled by default; flip on per-environment when the service-account
# client + vault secret are provisioned.
keycloak:
admin:
enabled: ${ORDINIS_KEYCLOAK_ADMIN_ENABLED:false}
url: ${ORDINIS_KEYCLOAK_ADMIN_URL:}
realm: ${ORDINIS_KEYCLOAK_ADMIN_REALM:nstart}
client-id: ${ORDINIS_KEYCLOAK_ADMIN_CLIENT_ID:}
client-secret: ${ORDINIS_KEYCLOAK_ADMIN_CLIENT_SECRET:}
@@ -0,0 +1,124 @@
package cloud.nstart.terravault.ordinis.domain.userdisplay;
import jakarta.persistence.Column;
import jakarta.persistence.Entity;
import jakarta.persistence.EnumType;
import jakarta.persistence.Enumerated;
import jakarta.persistence.Id;
import jakarta.persistence.Table;
import org.hibernate.annotations.UpdateTimestamp;
import java.time.OffsetDateTime;
/**
* Persistent cache entry for JWT sub → display info resolve. See
* migration 0023-user-display-cache.xml for the table contract and
* the multi-tier resolve strategy.
*
* <p>Source semantics:
* <ul>
* <li>{@link Source#JWT_CAPTURE} — written by JwtUserCaptureFilter
* on every authenticated request. Highest accuracy (claims as
* the user actually saw them).</li>
* <li>{@link Source#KEYCLOAK_SYNC} — written by the periodic
* Phase 3 scheduled bulk sync from Keycloak Admin API.</li>
* <li>{@link Source#KEYCLOAK_ON_DEMAND} — written when a cache miss
* triggers a one-shot Keycloak Admin lookup. Phase 2.</li>
* </ul>
*/
@Entity
@Table(name = "user_display_cache")
public class UserDisplayCacheEntry {
@Id
@Column(name = "sub", nullable = false, length = 64)
private String sub;
@Column(name = "preferred_username", length = 255)
private String preferredUsername;
@Column(name = "name", length = 255)
private String name;
@Column(name = "email", length = 320)
private String email;
@Enumerated(EnumType.STRING)
@Column(name = "source", nullable = false, length = 32)
private Source source = Source.JWT_CAPTURE;
@Column(name = "synced_at", nullable = false)
private OffsetDateTime syncedAt;
@UpdateTimestamp
@Column(name = "updated_at", nullable = false)
private OffsetDateTime updatedAt;
protected UserDisplayCacheEntry() {
// JPA
}
public UserDisplayCacheEntry(String sub, String preferredUsername, String name,
String email, Source source, OffsetDateTime syncedAt) {
this.sub = sub;
this.preferredUsername = preferredUsername;
this.name = name;
this.email = email;
this.source = source;
this.syncedAt = syncedAt;
}
public String getSub() {
return sub;
}
public String getPreferredUsername() {
return preferredUsername;
}
public String getName() {
return name;
}
public String getEmail() {
return email;
}
public Source getSource() {
return source;
}
public OffsetDateTime getSyncedAt() {
return syncedAt;
}
public OffsetDateTime getUpdatedAt() {
return updatedAt;
}
/**
* Idempotent merge from a fresh capture / sync. Updates fields that
* have a new non-null value; leaves existing data otherwise (e.g.
* Keycloak sync may have richer email than JWT capture, don't blank).
*/
public void mergeFrom(String preferredUsername, String name, String email,
Source source, OffsetDateTime syncedAt) {
if (preferredUsername != null && !preferredUsername.isBlank()) {
this.preferredUsername = preferredUsername;
}
if (name != null && !name.isBlank()) {
this.name = name;
}
if (email != null && !email.isBlank()) {
this.email = email;
}
this.source = source;
this.syncedAt = syncedAt;
}
public enum Source {
JWT_CAPTURE,
KEYCLOAK_SYNC,
KEYCLOAK_ON_DEMAND
}
}
@@ -0,0 +1,9 @@
package cloud.nstart.terravault.ordinis.domain.userdisplay;
import org.springframework.data.jpa.repository.JpaRepository;
public interface UserDisplayCacheRepository extends JpaRepository<UserDisplayCacheEntry, String> {
// PK is sub (String). Built-in findById(sub), save, etc. covers Phase 1+2.
// Phase 3 (scheduled bulk sync) will add `findAllBySyncedAtBefore(...)`
// and `findAllByOrderBySyncedAtAsc(Pageable)` for stale-first refresh.
}
@@ -0,0 +1,81 @@
<?xml version="1.0" encoding="UTF-8"?>
<databaseChangeLog
xmlns="http://www.liquibase.org/xml/ns/dbchangelog"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://www.liquibase.org/xml/ns/dbchangelog
http://www.liquibase.org/xml/ns/dbchangelog/dbchangelog-4.27.xsd">
<!--
User display cache (sub → username/email/name) — persistent backing store
for UserDisplayService.
Цель: остановить bleeding "ток мой UUID отображается" в audit/reviews/
history/changelog. Раньше UserDisplayService держал данные in-memory →
pod restart терял всё → users которые редко логинились показывались
как UUID.
Three-tier resolve chain (после wire-up в UserDisplayService):
1. JVM ConcurrentHashMap (hot cache, ~µs)
2. user_display_cache table (этот migration, ~ms)
3. Keycloak Admin API on-demand (~10-50ms, only if not in DB)
+ nightly scheduled bulk sync to refill (Phase 3).
Decisions:
- PK = sub (UUID-like Keycloak user id). Match'ит JWT sub claim
directly — нет surrogate id.
- source ENUM текстом ('JWT_CAPTURE' | 'KEYCLOAK_SYNC' |
'KEYCLOAK_ON_DEMAND') — Hibernate @Enumerated(STRING) уrok из 0020.
- synced_at = когда последний раз обновили из источника. updated_at =
row mtime (Hibernate @UpdateTimestamp). Разные потому что
KEYCLOAK_SYNC bulk job обновляет каждые 6h даже если данные не
изменились.
- Soft cap on rows не нужен — Keycloak realm пока ~50 users, table
будет ≤ 1k entries обозримо.
-->
<changeSet id="0023-1-user-display-cache" author="ordinis">
<comment>user_display_cache — persistent sub → display info, three-tier resolve</comment>
<createTable tableName="user_display_cache">
<column name="sub" type="VARCHAR(64)">
<constraints primaryKey="true" nullable="false"/>
</column>
<column name="preferred_username" type="VARCHAR(255)">
<constraints nullable="true"/>
</column>
<column name="name" type="VARCHAR(255)">
<constraints nullable="true"/>
</column>
<column name="email" type="VARCHAR(320)"><!-- RFC 5321 max -->
<constraints nullable="true"/>
</column>
<column name="source" type="VARCHAR(32)" defaultValue="JWT_CAPTURE">
<constraints nullable="false"/>
</column>
<column name="synced_at" type="TIMESTAMP WITH TIME ZONE" defaultValueComputed="NOW()">
<constraints nullable="false"/>
</column>
<column name="updated_at" type="TIMESTAMP WITH TIME ZONE" defaultValueComputed="NOW()">
<constraints nullable="false"/>
</column>
</createTable>
</changeSet>
<changeSet id="0023-2-user-display-cache-source-check" author="ordinis">
<comment>Enforce source enum values inline в DB чтобы insert с typo шумел</comment>
<sql>
ALTER TABLE user_display_cache
ADD CONSTRAINT user_display_cache_source_check
CHECK (source IN ('JWT_CAPTURE', 'KEYCLOAK_SYNC', 'KEYCLOAK_ON_DEMAND'));
</sql>
<rollback>
ALTER TABLE user_display_cache DROP CONSTRAINT user_display_cache_source_check;
</rollback>
</changeSet>
<changeSet id="0023-3-user-display-cache-synced-index" author="ordinis">
<comment>Index for Phase 3 nightly sync — pull stale entries by synced_at</comment>
<createIndex tableName="user_display_cache" indexName="idx_user_display_synced_at">
<column name="synced_at"/>
</createIndex>
</changeSet>
</databaseChangeLog>
@@ -32,5 +32,6 @@
<include file="changes/0020-fix-draft-status-case.xml" relativeToChangelogFile="true"/> <include file="changes/0020-fix-draft-status-case.xml" relativeToChangelogFile="true"/>
<include file="changes/0021-notifications.xml" relativeToChangelogFile="true"/> <include file="changes/0021-notifications.xml" relativeToChangelogFile="true"/>
<include file="changes/0022-schema-workflow.xml" relativeToChangelogFile="true"/> <include file="changes/0022-schema-workflow.xml" relativeToChangelogFile="true"/>
<include file="changes/0023-user-display-cache.xml" relativeToChangelogFile="true"/>
</databaseChangeLog> </databaseChangeLog>
@@ -0,0 +1,171 @@
package cloud.nstart.terravault.ordinis.restapi.service;
import com.fasterxml.jackson.annotation.JsonIgnoreProperties;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.beans.factory.annotation.Value;
import org.springframework.boot.autoconfigure.condition.ConditionalOnProperty;
import org.springframework.http.HttpHeaders;
import org.springframework.http.HttpStatus;
import org.springframework.http.MediaType;
import org.springframework.stereotype.Service;
import org.springframework.util.LinkedMultiValueMap;
import org.springframework.util.MultiValueMap;
import org.springframework.web.client.HttpClientErrorException;
import org.springframework.web.client.RestClient;
import org.springframework.web.util.UriComponentsBuilder;
import java.time.Duration;
import java.time.Instant;
import java.util.Optional;
/**
* Phase 2: resolve {@code sub} via Keycloak Admin REST API. Activated when
* {@code ordinis.keycloak.admin.enabled=true} (and required env vars set).
*
* <p>Auth flow: client_credentials grant against
* {@code {issuer}/protocol/openid-connect/token}. The service-account
* client must have realm role {@code view-users} (composite of
* {@code realm-management/view-users}).
*
* <p>Token caching: the access_token from Keycloak typically lives
* 60-300s. We cache it minus 30s safety margin and refresh on miss.
*
* <p>Lookup uses {@code GET /admin/realms/{realm}/users/{sub}} which
* returns 404 cleanly when the sub is unknown. Network/5xx are surfaced
* as RuntimeException so {@link UserDisplayService} can log and skip.
*/
@Service
@ConditionalOnProperty(name = "ordinis.keycloak.admin.enabled", havingValue = "true")
public class KeycloakAdminUserResolver implements KeycloakUserResolver {
private static final Logger log = LoggerFactory.getLogger(KeycloakAdminUserResolver.class);
private static final Duration TOKEN_REFRESH_MARGIN = Duration.ofSeconds(30);
private final String adminBaseUrl;
private final String realm;
private final String clientId;
private final String clientSecret;
private final RestClient http;
// Token cache — Keycloak access_token + expiry. Access is single-threaded
// in practice (one resolver call at a time); volatile is enough.
private volatile String cachedToken;
private volatile Instant cachedTokenExpiresAt = Instant.EPOCH;
public KeycloakAdminUserResolver(
@Value("${ordinis.keycloak.admin.url}") String adminBaseUrl,
@Value("${ordinis.keycloak.admin.realm}") String realm,
@Value("${ordinis.keycloak.admin.client-id}") String clientId,
@Value("${ordinis.keycloak.admin.client-secret}") String clientSecret) {
this.adminBaseUrl = trimTrailingSlash(adminBaseUrl);
this.realm = realm;
this.clientId = clientId;
this.clientSecret = clientSecret;
this.http = RestClient.builder()
.baseUrl(this.adminBaseUrl)
.build();
log.info("KeycloakAdminUserResolver active — base={} realm={} client={}",
this.adminBaseUrl, realm, clientId);
}
@Override
public Optional<KeycloakUser> findById(String sub) {
if (sub == null || sub.isBlank()) return Optional.empty();
String token = currentToken();
String url = UriComponentsBuilder.fromHttpUrl(adminBaseUrl)
.pathSegment("admin", "realms", realm, "users", sub)
.toUriString();
try {
KeycloakUserDto dto = http.get()
.uri(url)
.header(HttpHeaders.AUTHORIZATION, "Bearer " + token)
.retrieve()
.body(KeycloakUserDto.class);
if (dto == null || dto.id == null) return Optional.empty();
String name = composeName(dto.firstName, dto.lastName);
return Optional.of(new KeycloakUser(
dto.id,
dto.username,
name,
dto.email));
} catch (HttpClientErrorException.NotFound e) {
// Sub doesn't exist in this realm — definitive miss. Don't cache the
// negative; UserDisplayService.find returns Optional.empty and UserCell
// falls back to short UUID. Negative cache TTL would help latency on
// bots/scanners but adds complexity not justified yet.
return Optional.empty();
} catch (HttpClientErrorException.Unauthorized e) {
// Cached token rejected — wipe and retry once. If still 401, surface.
cachedToken = null;
cachedTokenExpiresAt = Instant.EPOCH;
throw new RuntimeException("Keycloak admin auth rejected (token wiped, will retry)", e);
}
}
// --- token management ---
private synchronized String currentToken() {
if (cachedToken != null && Instant.now().isBefore(cachedTokenExpiresAt)) {
return cachedToken;
}
return fetchAndCacheToken();
}
private String fetchAndCacheToken() {
String tokenUrl = UriComponentsBuilder.fromHttpUrl(adminBaseUrl)
.pathSegment("realms", realm, "protocol", "openid-connect", "token")
.toUriString();
MultiValueMap<String, String> form = new LinkedMultiValueMap<>();
form.add("grant_type", "client_credentials");
form.add("client_id", clientId);
form.add("client_secret", clientSecret);
TokenResponse tr = http.post()
.uri(tokenUrl)
.contentType(MediaType.APPLICATION_FORM_URLENCODED)
.body(form)
.retrieve()
.onStatus(s -> s.is4xxClientError() || s.is5xxServerError(), (req, res) -> {
throw new RuntimeException("Keycloak admin token request failed: HTTP "
+ res.getStatusCode() + " — check client_id/secret + realm");
})
.body(TokenResponse.class);
if (tr == null || tr.accessToken == null) {
throw new RuntimeException("Keycloak admin token response missing access_token");
}
cachedToken = tr.accessToken;
int expiresIn = tr.expiresIn != null ? tr.expiresIn : 60;
cachedTokenExpiresAt = Instant.now()
.plusSeconds(expiresIn)
.minus(TOKEN_REFRESH_MARGIN);
return cachedToken;
}
// --- helpers ---
private static String trimTrailingSlash(String s) {
return (s != null && s.endsWith("/")) ? s.substring(0, s.length() - 1) : s;
}
private static String composeName(String first, String last) {
String f = (first == null || first.isBlank()) ? "" : first.trim();
String l = (last == null || last.isBlank()) ? "" : last.trim();
String joined = (f + " " + l).trim();
return joined.isEmpty() ? null : joined;
}
// --- DTOs (Keycloak 26 admin API) ---
@JsonIgnoreProperties(ignoreUnknown = true)
private record TokenResponse(
@com.fasterxml.jackson.annotation.JsonProperty("access_token") String accessToken,
@com.fasterxml.jackson.annotation.JsonProperty("expires_in") Integer expiresIn) {}
@JsonIgnoreProperties(ignoreUnknown = true)
private record KeycloakUserDto(
String id,
String username,
String firstName,
String lastName,
String email) {}
}
@@ -0,0 +1,24 @@
package cloud.nstart.terravault.ordinis.restapi.service;
import java.util.Optional;
/**
* Phase 2 contract: resolve sub → user info via Keycloak Admin API. The
* implementation (Phase 2) will be an HTTP client using a service-account
* client + role {@code view-users} on realm {@code nstart}.
*
* <p>{@link UserDisplayService} consumes this as an optional dependency —
* if no bean exists in the context, the third resolve tier is skipped and
* UserCell falls back to short UUID. This keeps the Phase 1 (DB cache only)
* deployment runnable without Keycloak admin creds.
*
* <p>Implementations must throw {@link RuntimeException} on transient errors
* (network, 5xx) so the caller can log and continue. Return {@link Optional#empty()}
* for definitive 404 — sub does not exist in the realm.
*/
public interface KeycloakUserResolver {
Optional<KeycloakUser> findById(String sub);
record KeycloakUser(String sub, String preferredUsername, String name, String email) {}
}
@@ -1,69 +1,165 @@
package cloud.nstart.terravault.ordinis.restapi.service; package cloud.nstart.terravault.ordinis.restapi.service;
import cloud.nstart.terravault.ordinis.domain.userdisplay.UserDisplayCacheEntry;
import cloud.nstart.terravault.ordinis.domain.userdisplay.UserDisplayCacheEntry.Source;
import cloud.nstart.terravault.ordinis.domain.userdisplay.UserDisplayCacheRepository;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Service; import org.springframework.stereotype.Service;
import org.springframework.transaction.annotation.Transactional;
import java.time.OffsetDateTime; import java.time.OffsetDateTime;
import java.util.Optional; import java.util.Optional;
import java.util.concurrent.ConcurrentHashMap; import java.util.concurrent.ConcurrentHashMap;
/** /**
* In-memory кэш user display info (sub → preferredUsername / name / email), * Three-tier resolve sub → display info:
* заполняемый из JWT claims на каждом authenticated запросе через * <ol>
* {@link cloud.nstart.terravault.ordinis.restapi.auth.JwtUserCaptureFilter}. * <li>JVM ConcurrentHashMap hot cache (~µs).</li>
* <li>{@code user_display_cache} Postgres table (~ms, persistent across pod restart).</li>
* <li>Keycloak Admin API on-demand fallback (~10-50ms, Phase 2 — wired
* via optional {@link KeycloakUserResolver} bean).</li>
* </ol>
* *
* <p><b>Зачем:</b> audit log + record drafts хранят actor как JWT * <p>JwtUserCaptureFilter calls {@link #put} on every authenticated request
* {@code sub} UUID. Frontend (UserCell в audit / reviews) показывает short * → updates both hot cache and DB (idempotent merge).
* UUID + tooltip — плохой UX. Этот сервис позволяет резолвить sub →
* читаемое имя для display, не запрашивая Keycloak Admin API на каждое
* отображение.
* *
* <p><b>Trade-offs:</b> * <p>Phase 3 (TBD): scheduled bulk sync from Keycloak will call
* <ul> * {@link #upsertFromKeycloak} for every realm user every 6h to refill
* <li>In-memory — пропадает на pod restart. Постепенно заполняется * stale entries.
* обратно по мере того как users делают запросы. Для display это OK.</li>
* <li>Без Keycloak Admin API — не видим users которые ни разу не
* логинились в Ordinis. Display fallback → short UUID.</li>
* <li>Stale: если user сменит preferred_username в Keycloak, кэш
* обновится только на следующий его login + request.</li>
* </ul>
* *
* <p>Альтернатива (на будущее): миграция 0023 user_display_cache table * <p>Hot cache reads from DB lazily on miss — no warm-up at startup
* + periodic Keycloak sync. Сделать когда станет реально нужно * (would block readiness probe + Keycloak might be cold). First request
* (пока display name никто не редактирует постоянно). * for each sub after restart pays a single DB hit.
*/ */
@Service @Service
public class UserDisplayService { public class UserDisplayService {
/** Cap чтобы не держать миллионы entries в RAM при abuse. ~100 байт per private static final Logger log = LoggerFactory.getLogger(UserDisplayService.class);
* entry → 10k = 1MB. Eviction policy — LRU не нужен, hash без bound — fine. */
private static final int SOFT_CAP = 10_000;
private final ConcurrentHashMap<String, UserDisplayInfo> cache = new ConcurrentHashMap<>();
/** /**
* Обновляет / создаёт запись для {@code sub}. Idempotent. Вызывается * Cap чтобы не держать миллионы entries в RAM при abuse. ~100 байт per
* из {@code JwtUserCaptureFilter} на каждом authenticated request — * entry → 10k = 1MB. Eviction policy — LRU не нужен, hash без bound — fine.
* стоимость ~O(1) hash put. * DB table сама разрастается без cap (см. migration 0023 design notes).
*/ */
private static final int HOT_CACHE_CAP = 10_000;
private final ConcurrentHashMap<String, UserDisplayInfo> hotCache = new ConcurrentHashMap<>();
private final UserDisplayCacheRepository repository;
private final KeycloakUserResolver keycloakResolver;
public UserDisplayService(UserDisplayCacheRepository repository,
@Autowired(required = false) KeycloakUserResolver keycloakResolver) {
this.repository = repository;
this.keycloakResolver = keycloakResolver;
if (keycloakResolver == null) {
log.info("UserDisplayService: KeycloakUserResolver not wired — three-tier resolve "
+ "will fall back to short-UUID for users that never made an authenticated "
+ "request to ordinis.");
}
}
/**
* Capture from JWT (called per-request by JwtUserCaptureFilter). Cheap and
* idempotent. Writes to hot cache + DB. Source is always JWT_CAPTURE here —
* KEYCLOAK_SYNC / KEYCLOAK_ON_DEMAND have their own write paths.
*/
@Transactional
public void put(String sub, String preferredUsername, String name, String email) { public void put(String sub, String preferredUsername, String name, String email) {
if (sub == null || sub.isBlank()) return; if (sub == null || sub.isBlank()) return;
if (cache.size() >= SOFT_CAP) return; // soft cap, don't grow unbounded upsert(sub, preferredUsername, name, email, Source.JWT_CAPTURE);
cache.put(sub, new UserDisplayInfo(
sub,
nullIfBlank(preferredUsername),
nullIfBlank(name),
nullIfBlank(email),
OffsetDateTime.now()));
} }
/**
* Three-tier lookup. Returns empty only if sub absent in all three tiers
* AND Keycloak is unreachable (transient — caller's UserCell will fall
* back to short UUID, retry next time it renders).
*/
@Transactional
public Optional<UserDisplayInfo> find(String sub) { public Optional<UserDisplayInfo> find(String sub) {
if (sub == null || sub.isBlank()) return Optional.empty(); if (sub == null || sub.isBlank()) return Optional.empty();
return Optional.ofNullable(cache.get(sub));
// Tier 1: hot cache
UserDisplayInfo hot = hotCache.get(sub);
if (hot != null) return Optional.of(hot);
// Tier 2: DB
Optional<UserDisplayCacheEntry> dbHit = repository.findById(sub);
if (dbHit.isPresent()) {
UserDisplayInfo info = toInfo(dbHit.get());
cacheHot(sub, info);
return Optional.of(info);
}
// Tier 3: Keycloak Admin on-demand (Phase 2)
if (keycloakResolver != null) {
try {
Optional<KeycloakUserResolver.KeycloakUser> kc = keycloakResolver.findById(sub);
if (kc.isPresent()) {
KeycloakUserResolver.KeycloakUser u = kc.get();
UserDisplayInfo info = upsert(sub, u.preferredUsername(), u.name(), u.email(),
Source.KEYCLOAK_ON_DEMAND);
return Optional.of(info);
}
} catch (RuntimeException e) {
log.warn("Keycloak on-demand lookup failed for sub={}: {}", sub, e.getMessage());
}
}
return Optional.empty();
} }
/** Stats для observability — сколько users закэшировано. */ /**
public int size() { * Phase 3 entry point — scheduled sync writes every realm user with
return cache.size(); * Source.KEYCLOAK_SYNC. Same idempotent merge as {@link #put}.
*/
@Transactional
public void upsertFromKeycloak(String sub, String preferredUsername, String name, String email) {
if (sub == null || sub.isBlank()) return;
upsert(sub, preferredUsername, name, email, Source.KEYCLOAK_SYNC);
}
/** Stats для observability. */
public int hotCacheSize() {
return hotCache.size();
}
// --- internals ---
private UserDisplayInfo upsert(String sub, String preferredUsername, String name,
String email, Source source) {
OffsetDateTime now = OffsetDateTime.now();
UserDisplayCacheEntry entry = repository.findById(sub).orElse(null);
if (entry == null) {
entry = new UserDisplayCacheEntry(sub,
nullIfBlank(preferredUsername),
nullIfBlank(name),
nullIfBlank(email),
source,
now);
} else {
entry.mergeFrom(preferredUsername, name, email, source, now);
}
repository.save(entry);
UserDisplayInfo info = toInfo(entry);
cacheHot(sub, info);
return info;
}
private void cacheHot(String sub, UserDisplayInfo info) {
if (hotCache.size() < HOT_CACHE_CAP) {
hotCache.put(sub, info);
}
}
private static UserDisplayInfo toInfo(UserDisplayCacheEntry e) {
return new UserDisplayInfo(
e.getSub(),
e.getPreferredUsername(),
e.getName(),
e.getEmail(),
e.getUpdatedAt());
} }
private static String nullIfBlank(String s) { private static String nullIfBlank(String s) {