Technical Knowledge Accumulated from Real-World Production Operations
The GIIP Production Knowledge Base shares practical knowledge gained from real customer support, database migrations, cloud operations, performance tuning, and incident response, sanitized of confidential data and thoroughly re-verified.
Every document specifies target products, verified versions, required permissions, execution impact, and last verified dates. Read-only SQL queries can be directly copied and executed.
Keyword Search
33 documents
- SQL ServerMediumSQL to Check Statistics Last-Updated Time per Table in SQL ServerA read-only SQL script using sys.stats and STATS_DATE that lists the last-updated time and the number of rows modified since the last update, per table and statistic.2026-08-13
- SQL ServerRead-onlyHow to Check Transaction Log Usage on RDS for SQL ServerA read-only procedure that checks log usage with DBCC SQLPERF(LOGSPACE) and sys.dm_db_log_space_usage, and isolates the reason it is not being released using log_reuse_wait_desc.2026-08-13
- SQL ServerHighSQL to Check for Long-Running Open Transactions in SQL ServerA procedure using the sys.dm_tran_active_transactions family of DMVs to identify abandoned transactions, including start time, session, and the last SQL statement executed.2026-08-13
- SQL ServerMediumCauses and Checks for High CXPACKET Waits in SQL ServerA procedure for evaluating CXPACKET using sys.dm_os_wait_stats and sys.dm_os_waiting_tasks, distinguishing it from CXCONSUMER, and moving toward a review of the degree of parallelism.2026-08-13
- SQL ServerMediumHow to Check and Change MAXDOP and Cost Threshold for Parallelism in SQL ServerA procedure for checking and changing the degree of parallelism at the server, database, and query levels, along with the differences in scope and impact at each level.2026-08-13
- SQL ServerExpert Review RequiredHow to Check Whether CDC Log Scanning Has Stopped in SQL ServerA procedure for checking whether CDC's log scan is progressing, using sys.dm_cdc_log_scan_sessions, cdc.lsn_time_mapping, and the state of the capture job.2026-08-13
- SQL ServerExpert Review RequiredWhy MSrepl_commands Keeps Growing in SQL Server, and How to Check for Replication LagA procedure for isolating a backlog of commands in the distribution database from two angles: the state of the distribution agent, and the cleanup job / retention period.2026-08-13
- SQL ServerHighCauses and Checks for Msg 7356 on a Linked Server in SQL ServerMsg 7356 signals that column metadata disagreed between compile time and run time. This organizes how to isolate the cause and how to work around it using pass-through queries.2026-08-13
- SQL ServerMediumWhy TLS 1.2 Connections Fail with SQLNCLI10, and Migrating to MSOLEDBSQL 19This separates connection failures caused by SQLNCLI10's lack of TLS 1.2 support from those caused by MSOLEDBSQL 19's default changing to Encrypt=yes, and organizes what to check during migration.2026-08-13
- SQL ServerHighThe Difference Between DBCC SHRINKDATABASE and DBCC SHRINKFILE, and What to Check Before Running ThemOrganizes the difference between SHRINKDATABASE, which targets the whole database, and SHRINKFILE, which targets a single file; when to use TRUNCATEONLY; and what to check before running either.2026-08-13
- SQL ServerHighCauses and Countermeasures for Disk I/O Performance Degradation After Migrating an On-Premises SQL Server to AWSOrganizes the causes of disk I/O performance degradation due to structural differences between on-premises physical NVMe/SSD and AWS EBS, and four practical countermeasures: buffer pool maximization, data compression, file separation, and CDC log bloat response.2026-09-07
- Aurora MySQLHighHow to Investigate a Slow Query in Aurora MySQLA procedure for finding slow queries by starting with lower-risk checks, in the order: slow query log, PROCESSLIST, EXPLAIN, and digest aggregation. Also shows alternatives for environments where performance_schema is disabled.2026-08-13
- Aurora MySQLMediumWhy AWS DMS Reports Error 1032 and How to Investigate ItError 1032 in DMS CDC means the corresponding row doesn't exist on the target. This organizes how to read the control tables, how to check for primary key mismatches, and the available task-setting options.2026-08-13
- Aurora MySQLHighWhat to Check When Migrating from utf8mb3 to utf8mb4 in Aurora MySQLThe first thing that breaks when converting to utf8mb4 is index key length. This organizes the SQL for finding targets, how to choose a collation, the impact of conversion, and where the settings live.2026-08-13
- Aurora MySQLHighHow to Check and Configure binlog Retention in Aurora MySQLCheck binlog retention with `mysql.rds_show_configuration` and set it in hours via `mysql.rds_set_configuration`. `binlog_format` is a cluster parameter and changing it requires a restart.2026-08-13
- Aurora MySQLHighHow to Collect and Review Audit Logs in Aurora MySQLA procedure for enabling Aurora MySQL's Advanced Auditing via a cluster parameter and reviewing it through CloudWatch Logs or RDS log files. Also organizes how to narrow down what gets recorded.2026-08-13
- TiDBExpert Review RequiredChecklist for Migrating from TiDB to Aurora MySQLTiDB is MySQL-compatible, but not identical. This organizes the differences in primary keys, ID generation, transactions, statistics, and capacity into a pre-migration checklist with verification SQL.2026-08-13
- Database MigrationMediumHow to Plan a 3TB-Scale Database MigrationA 3TB-class migration plan is built in this order: measure actual size, choose a cutover method, measure duration with a trial run, and design verification and rollback. This is a procedure for setting the schedule from measured values rather than rule-of-thumb figures.2026-08-13
- AWSHighThe Difference Between One Large RDS Instance and Several Smaller InstancesA document organizing the technical differences between "one large instance" and "several smaller instances" for RDS sizing, and the CloudWatch metrics you should measure before deciding.2026-08-13
- AWSHighWhat to Check When Reviewing AWS Database CostsA checklist for reviewing AWS database spend in order of increasing impact: stop what is unused → right-size → storage → backups → non-production → commitments.2026-08-13
- InfrastructureLowWhat's the Difference Between EBS IOPS and Physical Disk IOPSA document organizing why EBS IOPS cannot be compared directly to physical disk IOPS, based on where the ceiling applies (volume vs. instance) and how to measure it.2026-08-13
- InfrastructureMediumRandom IOPS vs. Sequential IOPS, and How to Read Benchmark ResultsA document organizing why an IOPS figure cannot be compared without I/O size, queue depth, and read/write ratio, and how to build measurement conditions with fio.2026-08-13
- MonitoringLowWhat to Configure When Monitoring Servers and Databases 24/7A checklist, organized by layer, of what to monitor, how to think about thresholds, the escalation structure, and why synthetic monitoring is necessary when designing 24/7 monitoring.2026-08-13
- GIIPRead-onlyWhat is the difference between a coding agent and GIIP FDE Ops?Coding agents and operations services differ in their "unit of work." We compare the boundary between the two across nine dimensions, with commands you can run to check your own organization.2026-08-13
- AI OperationsRead-onlyWhat it takes to bring an AI-generated application to productionWe organize the remaining work between "the code works" and "it can run in production" into 14 items, with commands for detecting hardcoded keys and auditing dependency vulnerabilities.2026-08-13
- AI OperationsRead-onlyWhy AI automated execution needs approval and rollback, and how to design for itDesign elements for automated execution — snapshots, approval gates, dry-run separation, allowlists, idempotency, audit logs, staged rollout, and a kill switch — are summarized, along with where to draw the line on what can run without approval.2026-08-13
- AI OperationsRead-onlyHow to Reduce Model Costs with an AI Router, and Failover Design for External AI API OutagesThis article summarizes a design for reducing model costs through task-difficulty-based routing and quality gates, along with a failover design assuming multiple providers. It does not cover pricing or cost-reduction figures.2026-08-13
- Incident ResponseExpert Review RequiredWhat database incidents AI agents can handle, and what humans must decideWe split incident response into phases to separate what AI can own (detection, triage, limited initial response) from what requires human judgment (irreversible actions, shutdown decisions), and provide read-only triage queries.2026-08-13
- AI OperationsRead-onlyWhy GIIP AI Harness Excludes Query HintsWhy GIIP AI Harness excludes FORCE INDEX and join hints from SQL generation and tuning, and the steps to safely evaluate indexes marked for deletion.2026-09-06
- SQL ServerRead-onlyMissing N prefix on NVARCHAR literals silently mangles Japanese/Korean text and emoji without any errorAn NVARCHAR literal missing its N prefix is implicitly converted with no error, losing non-ASCII characters. This summarizes detection methods and how to prevent recurrence.2026-08-16
- AI OperationsLowA guard that detects contamination in AI-generated content, and an operating rule that never treats regeneration alone as "done"This explains an operating guard that mechanically detects known contamination patterns in AI-generated content and never calls a regeneration "done" until someone has directly checked the resulting body text.2026-08-16
- GPU InfrastructureRead-onlyGPU and Server Configuration for Enterprise Local Drawing AI — 20-User Reference CostEnterprise local drawing AI requires RTX PRO 6000 Blackwell 96GB ×2 (20-user recommended) as the standard, with GPU selection depending on VRAM capacity, model size, concurrent users, drawing page count, and response time targets — not just user count.2026-08-27
- AI OperationsRead-onlyWhat is TypeSafe AI Jev? — Fast "System One" Decision Model Features & ArchitectureUnlike text-generating LLMs, Jev is an ultra-fast System One decision model that returns structured choices and scores in a single pass. Designed by TypeSafe AI, it delivers 40-200x faster execution and 400x cost savings for agent routing and guardrails.2026-09-21
Browse by Category
SQL Server12From statistics, wait types, transactions, and log files to CDC, replication, linked servers, and TLS connections — practical checks for SQL Server production operations, with SQL included.Aurora MySQL5Covers the judgment calls Aurora MySQL operations require — slow query investigation, DMS errors, character set migration, binlog retention, and audit logging.TiDB1Covers TiDB's operational characteristics and the differences to check when migrating to a MySQL-compatible database.Database Migration1Covers migration planning for large-scale databases, choosing a cutover approach, and post-migration validation checkpoints.AWS2Covers RDS instance configuration, scale-up vs. scale-out decisions, and checkpoints for reviewing database costs.Infrastructure2Covers the basics needed to avoid misreading the numbers — the difference between EBS IOPS and physical disk IOPS, random vs. sequential IOPS, and NVMe configuration bottlenecks.Monitoring1Covers what to actually watch in 24/7 monitoring, how to set thresholds, and who should own first-line response.Incident Response1Covers the process from incident detection to recovery, what can be delegated to AI agents, and what still requires human judgment.AI Operations6Covers the design decisions required to bring AI into operations — approval and rollback, audit logs, model costs, and failover when an external API fails.GPU Infrastructure1Knowledge on GPU server configuration, power, colocation, and cost comparisons between local LLMs and commercial APIs — published progressively.GIIP1A technical explanation of the GIIP operations model — FDE Ops, FDE Box, how it differs from coding agents, and turning AI-built apps into services.
Browse by Symptom
Slow response or high CPU usage
- SQL to Check Statistics Last-Updated Time per Table in SQL Server
- Causes and Checks for High CXPACKET Waits in SQL Server
- How to Check and Change MAXDOP and Cost Threshold for Parallelism in SQL Server
- Why MSrepl_commands Keeps Growing in SQL Server, and How to Check for Replication Lag
- The Difference Between DBCC SHRINKDATABASE and DBCC SHRINKFILE, and What to Check Before Running Them
- How to Investigate a Slow Query in Aurora MySQL
- What to Check When Migrating from utf8mb3 to utf8mb4 in Aurora MySQL
- Checklist for Migrating from TiDB to Aurora MySQL
- The Difference Between One Large RDS Instance and Several Smaller Instances
- What's the Difference Between EBS IOPS and Physical Disk IOPS
- Random IOPS vs. Sequential IOPS, and How to Read Benchmark Results
- What to Configure When Monitoring Servers and Databases 24/7
- Why GIIP AI Harness Excludes Query Hints
Disk or log keeps growing
- How to Check Transaction Log Usage on RDS for SQL Server
- SQL to Check for Long-Running Open Transactions in SQL Server
- How to Check Whether CDC Log Scanning Has Stopped in SQL Server
- Why MSrepl_commands Keeps Growing in SQL Server, and How to Check for Replication Lag
- The Difference Between DBCC SHRINKDATABASE and DBCC SHRINKFILE, and What to Check Before Running Them
- Causes and Countermeasures for Disk I/O Performance Degradation After Migrating an On-Premises SQL Server to AWS
- How to Investigate a Slow Query in Aurora MySQL
- Why AWS DMS Reports Error 1032 and How to Investigate It
- How to Check and Configure binlog Retention in Aurora MySQL
- How to Collect and Review Audit Logs in Aurora MySQL
- How to Plan a 3TB-Scale Database Migration
- What to Check When Reviewing AWS Database Costs
- What's the Difference Between EBS IOPS and Physical Disk IOPS
- Random IOPS vs. Sequential IOPS, and How to Read Benchmark Results
- What to Configure When Monitoring Servers and Databases 24/7
Can't connect or getting errors
Considering migration or a configuration change
- Why TLS 1.2 Connections Fail with SQLNCLI10, and Migrating to MSOLEDBSQL 19
- Causes and Countermeasures for Disk I/O Performance Degradation After Migrating an On-Premises SQL Server to AWS
- What to Check When Migrating from utf8mb3 to utf8mb4 in Aurora MySQL
- Checklist for Migrating from TiDB to Aurora MySQL
- How to Plan a 3TB-Scale Database Migration
- The Difference Between One Large RDS Instance and Several Smaller Instances
- What to Check When Reviewing AWS Database Costs
- Missing N prefix on NVARCHAR literals silently mangles Japanese/Korean text and emoji without any error
Building monitoring and incident response
- How to Check Transaction Log Usage on RDS for SQL Server
- SQL to Check for Long-Running Open Transactions in SQL Server
- Causes and Checks for High CXPACKET Waits in SQL Server
- How to Check Whether CDC Log Scanning Has Stopped in SQL Server
- Why MSrepl_commands Keeps Growing in SQL Server, and How to Check for Replication Lag
- Causes and Checks for Msg 7356 on a Linked Server in SQL Server
- Why AWS DMS Reports Error 1032 and How to Investigate It
- How to Collect and Review Audit Logs in Aurora MySQL
- How to Plan a 3TB-Scale Database Migration
- What to Configure When Monitoring Servers and Databases 24/7
- What is the difference between a coding agent and GIIP FDE Ops?
- What it takes to bring an AI-generated application to production
- Why AI automated execution needs approval and rollback, and how to design for it
- How to Reduce Model Costs with an AI Router, and Failover Design for External AI API Outages
- What database incidents AI agents can handle, and what humans must decide
- Why GIIP AI Harness Excludes Query Hints
- Missing N prefix on NVARCHAR literals silently mangles Japanese/Korean text and emoji without any error
- A guard that detects contamination in AI-generated content, and an operating rule that never treats regeneration alone as "done"
- What is TypeSafe AI Jev? — Fast "System One" Decision Model Features & Architecture
Recently Updated Documents
TypeSafe AI Jev (System One Decision Model)What is TypeSafe AI Jev? — Fast "System One" Decision Model Features & ArchitectureUnlike text-generating LLMs, Jev is an ultra-fast System One decision model that returns structured choices and scores in a single pass. Designed by TypeSafe AI, it delivers 40-200x faster execution and 400x cost savings for agent routing and guardrails.SQL ServerCauses and Countermeasures for Disk I/O Performance Degradation After Migrating an On-Premises SQL Server to AWSOrganizes the causes of disk I/O performance degradation due to structural differences between on-premises physical NVMe/SSD and AWS EBS, and four practical countermeasures: buffer pool maximization, data compression, file separation, and CDC log bloat response.GIIP AI HarnessWhy GIIP AI Harness Excludes Query HintsWhy GIIP AI Harness excludes FORCE INDEX and join hints from SQL generation and tuning, and the steps to safely evaluate indexes marked for deletion.企業向けローカルAIインフラGPU and Server Configuration for Enterprise Local Drawing AI — 20-User Reference CostEnterprise local drawing AI requires RTX PRO 6000 Blackwell 96GB ×2 (20-user recommended) as the standard, with GPU selection depending on VRAM capacity, model size, concurrent users, drawing page count, and response time targets — not just user count.SQL ServerMissing N prefix on NVARCHAR literals silently mangles Japanese/Korean text and emoji without any errorAn NVARCHAR literal missing its N prefix is implicitly converted with no error, losing non-ASCII characters. This summarizes detection methods and how to prevent recurrence.AI運用（自動生成コンテンツ全般）A guard that detects contamination in AI-generated content, and an operating rule that never treats regeneration alone as "done"This explains an operating guard that mechanically detects known contamination patterns in AI-generated content and never calls a regeneration "done" until someone has directly checked the resulting body text.
Popular Documents
Hand-picked by our editorial team based on broad relevance and high foundational value, rather than raw traffic estimates.
SQL ServerSQL to Check Statistics Last-Updated Time per Table in SQL ServerA read-only SQL script using sys.stats and STATS_DATE that lists the last-updated time and the number of rows modified since the last update, per table and statistic.SQL ServerSQL to Check for Long-Running Open Transactions in SQL ServerA procedure using the sys.dm_tran_active_transactions family of DMVs to identify abandoned transactions, including start time, session, and the last SQL statement executed.Aurora MySQLWhy AWS DMS Reports Error 1032 and How to Investigate ItError 1032 in DMS CDC means the corresponding row doesn't exist on the target. This organizes how to read the control tables, how to check for primary key mismatches, and the available task-setting options.ストレージRandom IOPS vs. Sequential IOPS, and How to Read Benchmark ResultsA document organizing why an IOPS figure cannot be compared without I/O size, queue depth, and read/write ratio, and how to build measurement conditions with fio.AWS RDSThe Difference Between One Large RDS Instance and Several Smaller InstancesA document organizing the technical differences between "one large instance" and "several smaller instances" for RDS sizing, and the CloudWatch metrics you should measure before deciding.GIIP FDE OpsWhat is the difference between a coding agent and GIIP FDE Ops?Coding agents and operations services differ in their "unit of work." We compare the boundary between the two across nine dimensions, with commands you can run to check your own organization.
Copy-and-Run SQL
Documents containing read-only SQL queries that do not alter data or configurations. Please verify risk labels and applicability conditions before executing any command.
SQL ServerSQL to Check Statistics Last-Updated Time per Table in SQL ServerA read-only SQL script using sys.stats and STATS_DATE that lists the last-updated time and the number of rows modified since the last update, per table and statistic.SQL ServerHow to Check Transaction Log Usage on RDS for SQL ServerA read-only procedure that checks log usage with DBCC SQLPERF(LOGSPACE) and sys.dm_db_log_space_usage, and isolates the reason it is not being released using log_reuse_wait_desc.SQL ServerSQL to Check for Long-Running Open Transactions in SQL ServerA procedure using the sys.dm_tran_active_transactions family of DMVs to identify abandoned transactions, including start time, session, and the last SQL statement executed.SQL ServerCauses and Checks for High CXPACKET Waits in SQL ServerA procedure for evaluating CXPACKET using sys.dm_os_wait_stats and sys.dm_os_waiting_tasks, distinguishing it from CXCONSUMER, and moving toward a review of the degree of parallelism.SQL ServerHow to Check and Change MAXDOP and Cost Threshold for Parallelism in SQL ServerA procedure for checking and changing the degree of parallelism at the server, database, and query levels, along with the differences in scope and impact at each level.SQL ServerHow to Check Whether CDC Log Scanning Has Stopped in SQL ServerA procedure for checking whether CDC's log scan is progressing, using sys.dm_cdc_log_scan_sessions, cdc.lsn_time_mapping, and the state of the capture job.
Database Migration
SQL ServerWhy TLS 1.2 Connections Fail with SQLNCLI10, and Migrating to MSOLEDBSQL 19This separates connection failures caused by SQLNCLI10's lack of TLS 1.2 support from those caused by MSOLEDBSQL 19's default changing to Encrypt=yes, and organizes what to check during migration.SQL ServerCauses and Countermeasures for Disk I/O Performance Degradation After Migrating an On-Premises SQL Server to AWSOrganizes the causes of disk I/O performance degradation due to structural differences between on-premises physical NVMe/SSD and AWS EBS, and four practical countermeasures: buffer pool maximization, data compression, file separation, and CDC log bloat response.Aurora MySQLWhat to Check When Migrating from utf8mb3 to utf8mb4 in Aurora MySQLThe first thing that breaks when converting to utf8mb4 is index key length. This organizes the SQL for finding targets, how to choose a collation, the impact of conversion, and where the settings live.TiDBChecklist for Migrating from TiDB to Aurora MySQLTiDB is MySQL-compatible, but not identical. This organizes the differences in primary keys, ID generation, transactions, statistics, and capacity into a pre-migration checklist with verification SQL.
AI Monitoring & Operations
AI運用What it takes to bring an AI-generated application to productionWe organize the remaining work between "the code works" and "it can run in production" into 14 items, with commands for detecting hardcoded keys and auditing dependency vulnerabilities.AI運用Why AI automated execution needs approval and rollback, and how to design for itDesign elements for automated execution — snapshots, approval gates, dry-run separation, allowlists, idempotency, audit logs, staged rollout, and a kill switch — are summarized, along with where to draw the line on what can run without approval.AI運用How to Reduce Model Costs with an AI Router, and Failover Design for External AI API OutagesThis article summarizes a design for reducing model costs through task-difficulty-based routing and quality gates, along with a failover design assuming multiple providers. It does not cover pricing or cost-reduction figures.GIIP AI HarnessWhy GIIP AI Harness Excludes Query HintsWhy GIIP AI Harness excludes FORCE INDEX and join hints from SQL generation and tuning, and the steps to safely evaluate indexes marked for deletion.
GIIP's Operations Model
Get a Free Assessment of Your Current DB & Infrastructure
Running these checks once is straightforward. The real challenge is running them continuously across multiple environments and responding immediately when anomalies occur.
Request DB & Infrastructure Assessment