{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"auto-random-generated-value-range","title":"AUTO_RANDOM generated-value range","keywords":["AUTO_RANDOM","auto random","MaxInt64","2^63","explicit ID","explicitly inserted ID"],"applicability":"Questions about whether AUTO_RANDOM(S, R) keeps every existing, migrated, generated, or explicitly inserted ID within a downstream numeric limit.","reference":"topics/auto-random-generated-value-range.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"ticdc-canal-json-watermark-by-sink","title":"TiCDC Canal-JSON WATERMARK by sink transport","keywords":["TiCDC","Canal-JSON","TIDB_WATERMARK","WATERMARK","Cloud Storage","S3","checkpoint-ts"],"applicability":"Questions about TIDB_WATERMARK meaning, filtering, frequency, or physical placement for MQ and Cloud Storage sinks.","reference":"topics/ticdc-canal-json-watermark-by-sink.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"ticdc-architecture-version-topology-boundaries","title":"TiCDC architecture, version, and deployment topology boundaries","keywords":["TiCDC","CDC version","new architecture","classic architecture","newarch","dedicated","shared","rep-cdc","Dataflow","checkpoint","ResolvedTS"],"applicability":"TiCDC diagnostics or questions that must distinguish the independent TiCDC version/build, source TiDB version, dedicated/shared deployment, Dataflow control-plane identity, and new/classic TiCDC kernel architecture.","reference":"topics/ticdc-architecture-version-topology-boundaries.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"tidb-low-replica-placement-policy","title":"TiDB low-replica Placement Policy","keywords":["Placement Policy","PLACEMENT POLICY","FOLLOWERS","replica count","single replica","two replicas","TiKV voting replica"],"applicability":"Questions asking how to configure or verify one or two TiKV voting replicas with Placement Policy, especially for TiDB 8.5.","reference":"topics/tidb-low-replica-placement-policy.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"tidb-resource-group-ru-sizing","title":"TiDB Resource Group RU sizing","keywords":["Resource Group","resource group","resource control","RU_PER_SEC","RU quota","RU budget","RU 配额","RU 预算","资源组","BURSTABLE","CALIBRATE RESOURCE"],"applicability":"Questions asking how to choose, recommend, or validate a Resource Group RU_PER_SEC value, especially when a TiDB Cloud example value might be mistaken for a production sizing recommendation.","reference":"topics/tidb-resource-group-ru-sizing.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"wetech-prepared-statement-boundary","title":"Wetech prepared-statement guidance boundary","keywords":["PREPARE","prepared statement","prepared SQL","precompiled statement","预编译"],"applicability":"Wetech questions where a reply might otherwise recommend PREPARE, prepared SQL, or a precompiled-statement workflow.","reference":"topics/wetech-prepared-statement-boundary.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"aws-public-endpoint-idle-timeout","title":"AWS Public Endpoint 340-second idle timeout","keywords":["Public Endpoint","public endpoint","公网地址","公网连接","DBeaver","5分钟","five minutes","340 seconds","340s","AWS Global Accelerator","TCP keep-alive","PrivateLink","Private Endpoint"],"applicability":"TiDB Cloud Starter or Essential on AWS where a SQL client connection through a Public Endpoint is terminated around five minutes, especially during a long-running query with no connection traffic.","reference":"topics/aws-public-endpoint-idle-timeout.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"tidb-batch-on-shard-column-semantics","title":"TiDB BATCH ON shard-column and actual batch-size semantics","keywords":["BATCH ON","batch on","非事务 DML","划分列","分批列","shard column","actual batch size","实际批次","create_time"],"applicability":"TiDB non-transactional BATCH DML questions about choosing the shard column, why actual affected rows can exceed LIMIT, or how to validate and tune a DELETE, UPDATE, INSERT, or REPLACE batch safely.","reference":"topics/tidb-batch-on-shard-column-semantics.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"tiflash-duplicate-mpp-task-dispatch","title":"TiFlash duplicate MPP task dispatch and task-is-already-registered failure","keywords":["task is already registered","Failed to register MPP Task","TiFlash MPP task register failed","duplicate MPP task dispatch","MPP task registration","MPP任务注册失败"],"applicability":"TiDB or TiDB Cloud tickets whose supplied error says TiFlash failed to register an MPP task because the task is already registered; not generic TiFlash failures, resource pressure, or unrelated Error 1105 reports.","reference":"topics/tiflash-duplicate-mpp-task-dispatch.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"tiflash-restart-query-impact-by-version","title":"TiFlash restart query impact and graceful shutdown version boundary","keywords":["restart TiFlash","TiFlash restart","重启 TiFlash","重启tiflash","TiFlash 重启","业务SQL会报错","graceful shutdown","graceful_wait_shutdown_timeout","v7.5.6"],"applicability":"Questions asking whether a planned TiFlash process or host restart can interrupt running SQL and how to reduce that impact, especially for v7.5.6 or another version before TiFlash graceful shutdown was introduced in v8.5.4.","reference":"topics/tiflash-restart-query-impact-by-version.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"wetech-cloud-max-tikv-memory-alert","title":"Wetech Cloud maximum TiKV node memory alert","keywords":["MaxTiKVMemoryUsageTooHigh","max memory utilization across TiKV nodes","TiKV node Max memory utilization alert","TiKV memory warning","TiKV memory alert","TiKV 内存告警","TiKV内存告警"],"applicability":"Only a TiDB Cloud ticket with a non-empty orgId whose question contains the exact Wetech-managed MaxTiKVMemoryUsageTooHigh name or maximum-TiKV-node alert text; that alert identity is sufficient even when Jira Account Alias is absent, and per-node peaks may be acquired later; never Wetech OP, APID, or an ordinary memory question without an alert identity.","reference":"topics/wetech-cloud-max-tikv-memory-alert.md","diagnosticScenario":"alert-firing-explanation"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"wetech-cloud-total-tikv-memory-alert","title":"Wetech Cloud total TiKV node memory alert","keywords":["TotalTiKVMemoryUsageTooHigh","total memory utilization across TiKV nodes","Total TiKV node memory utilization alert","TiKV memory utilization across cluster alert","TiKV memory warning","TiKV memory alert","TiKV 内存告警","TiKV内存告警"],"applicability":"Only a TiDB Cloud ticket with a non-empty orgId whose question contains the exact Wetech-managed TotalTiKVMemoryUsageTooHigh name or total-TiKV-cluster alert text; that alert identity is sufficient even when Jira Account Alias is absent, and the cluster-wide average may be acquired later; never Wetech OP, APID, or an ordinary memory question without an alert identity.","reference":"topics/wetech-cloud-total-tikv-memory-alert.md","diagnosticScenario":"alert-firing-explanation"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"wetech-cloud-ska-environment","title":"Wetech Cloud SKA environment context","keywords":["1372813089209061633"],"applicability":"Only tickets whose typed orgId is exactly 1372813089209061633. This identifies the Wetech TiDB Cloud SKA environment, which is primarily Dedicated; it is internal customer context and never changes the non-empty-orgId Cloud boundary into Wetech OP.","reference":"topics/wetech-cloud-ska-environment.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"dify-tidb-vector-integration","title":"Dify integration with TiDB Vector, Data Service, and MCP","keywords":["Dify","TiDB Vector","tidb_vector","VECTOR_STORE","Knowledge Pipeline","Data Service Dify","Dify MCP"],"applicability":"Questions asking whether or how Dify can use TiDB or TiDB Cloud, especially when the answer must distinguish Dify's native TiDB Vector knowledge-store backend from Data Service/OpenAPI tools and a separately hosted TiDB MCP Server.","reference":"topics/dify-tidb-vector-integration.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"aws-privatelink-endpoint-traffic-ownership","title":"AWS PrivateLink endpoint traffic ownership","keywords":["vpce-","VPC endpoint traffic","interface endpoint traffic","endpoint traffic","endpoint in use","endpoint unused","delete endpoint","删除 endpoint","端点流量"],"applicability":"Questions asking whether a named AWS interface VPC endpoint used for a TiDB Cloud Private Endpoint has traffic, remains in use, or can be deleted; not Public Endpoint connectivity, generic cluster traffic, or endpoint setup questions.","reference":"topics/aws-privatelink-endpoint-traffic-ownership.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"aws-privatelink-security-ip-ownership","title":"AWS PrivateLink security IP ownership","keywords":["Private Link 安全配置","PrivateLink 安全配置","收紧 Private Link 访问 IP 范围","收紧 PrivateLink 访问 IP 范围","Private Link allowlist","PrivateLink allowlist","Private Endpoint ENI IP"],"applicability":"Questions asking which IP address or CIDR belongs in a security rule for a TiDB Cloud AWS Private Endpoint, or how to narrow a PrivateLink allowlist; not Public Endpoint connectivity, cross-Region support, endpoint traffic or deletion decisions, or generic cluster network diagnostics.","reference":"topics/aws-privatelink-security-ip-ownership.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"aws-cross-region-private-endpoint-restriction","title":"AWS cross-region Private Endpoint restriction for TiDB Cloud Starter and Essential","keywords":["AWS PrivateLink","Private Endpoint","endpoint service","cross-region","跨区域","Frankfurt","eu-central-1","Hong Kong","ap-east-1"],"applicability":"Only TiDB Cloud Starter or Essential instances on AWS when the AWS interface endpoint and TiDB Cloud endpoint service are in different AWS Regions. The restriction does not establish behavior for TiDB Cloud Dedicated and does not apply to same-region Private Endpoint connections.","reference":"topics/aws-cross-region-private-endpoint-restriction.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"op-support-project-environment-boundary","title":"APID and NAID OP support project environment boundary","keywords":["APID","NAID","OP TiCDC","OP CDC"],"applicability":"Only tickets whose typed Jira projectKey is exactly APID or NAID. These projects are OP; the runtime classification remains the sole Provider owner, Wetech or apac-test is the only supported OP TiCDC observability route, and Cloud Dataflow changefeed metadata is not applicable.","reference":"topics/op-support-project-environment-boundary.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"wetech-regional-saas-shorthand","title":"Wetech Asia SaaS shorthand and customer confirmation","keywords":["亚太01","亚太02","亚太 01","亚太 02"],"applicability":"Only Wetech OP tickets routed to Wetech with explicit 亚太01/亚太 01 or 亚太02/亚太 02: default respectively to the exact current inventory candidate asia-prod/SaaS01 or asia-prod/SaaS02 and diagnose before confirmation. Explicit different business/full names take precedence; conflicting regions or absent candidates remain unresolved. Select this topic and preserve the assumed mapping in the scenario reason; reconciliation must state the assumed customer-readable cluster and ask the customer to confirm it, retaining human_required, clarificationSimulation.required=true and focused requestedContext. No default for bare region, other numbers, accounts or Cloud.","reference":"topics/wetech-regional-saas-shorthand.md"}
{"schemaVersion":"auto-oncall-product-knowledge-index-v1","id":"tidb-top-sql-other-aggregation","title":"TiDB Top SQL Other is the long-tail query aggregate","keywords":["Top SQL","TopSQL","Other Statements","Others","长尾 query","长尾查询","长尾 SQL"],"applicability":"Interpreting the Other/Others/Other Statements bucket in TiDB Dashboard Top SQL or an explicitly equivalent query-tail summary: it aggregates all queries outside the displayed Top N, not one unidentified query. Counts and resource totals belong to the combined tail. Do not apply to unrelated Other labels.","reference":"topics/tidb-top-sql-other-aggregation.md"}