SQL SERVER CONSULTING & DBA SERVICES — SINCE 1990
Expert SQL Server Consulting | Remote DBA and Performance Tuning since 1990
35 years running SQL Server in mission-critical environments — ERP, core financial, e-commerce and national defense.
24/7 Remote DBA, Performance Tuning, Health Check and Emergency Support.
1990
In continuous operation since — 35 years of enterprise databases
1,000+
Production database servers currently operated by HTI
596+
Production database incidents resolved by our 24/7 NOC
99.99%
SLA sustained with Always On AG in financial customers
100%
Satisfaction rate across 596+ engagements
THE DIFFERENCE
HTI has been running SQL Server since its earliest versions in Brazilian corporate Windows environments. We are not generalists with a Microsoft certification — we are a team dedicated to production databases 24/7.
The depth of a team that has already seen — and fixed — nearly 600 incidents in production.
In-house team in a controlled room, with senior SQL Server DBAs on call in Portuguese and English.
Database consulting since 1990. We support SQL Server in ERP, core financial, e-commerce and national defense.
Always On, TDE, Query Store, Extended Events, well-sized tempdb, managed CU/GDR patching — no improvisation.
SQL SERVER CONSULTING SERVICES
From emergency diagnostics to continuous operations — we cover the full spectrum of SQL Server DBA services in production.
End-to-end SQL Server tuning — Query Store, wait statistics, stale statistics, parameterized plans, parameter sniffing, max server memory, tempdb and MAXDOP. Analysis of CPU, PAGEIOLATCH, PAGELATCH and lock waits — from symptom to root cause.
Audit of logins and permissions, TDE, Always Encrypted, Row-Level Security, Dynamic Data Masking, sysadmin review, LGPD/GDPR compliance. Hardening aligned with CIS Benchmarks and Microsoft Security Baseline — with an executable remediation plan.
Always On Availability Groups (synchronous and asynchronous), Failover Cluster Instances, Log Shipping, legacy Database Mirroring, geographically distributed replicas. Design, deployment and periodic failover testing.
Full SQL Server operations — 24/7 monitoring, incident response, proactive tuning, capacity planning, CU/GDR patch management. A team of senior SQL Server DBAs without the cost of in-house hiring.
Zero-downtime migration between versions (2014 → 2016 → 2019 → 2022), moves to Azure SQL Managed Instance, RDS, or SQL Server → PostgreSQL exits (licensing reduction). Compatibility level, DMA, regression testing.
SQL Server down? TempDB full? Deadlocks in production? 24/7 emergency support — 30-minute SLA in business hours, no prior contract required. Reachable via WhatsApp or phone.
Schema review, partitioning, indexing strategy (rowstore + columnstore), filegroups, tempdb sizing, MAXDOP and Cost Threshold for Parallelism. Complete technical audit with a roadmap prioritized by impact.
Azure SQL Database, Managed Instance, SQL Server on VM, Amazon RDS for SQL Server — configuration, right-sizing, licensing reduction (BYOL vs license-included) and database FinOps.
Full diagnostic: performance, security, HA/DR, backups, tempdb, waits, missing/duplicate indexes, instance configuration. Deliverable: report with findings ranked by severity and prioritized action plan.
TECHNICAL GUIDE
The main performance adjustments an experienced SQL Server DBA makes — and that the product's default install ignores.
Without max server memory configured, SQL Server consumes almost all available RAM and starves the OS. Rule of thumb: leave 4 GB + 1 GB for every 4 GB above 16 GB to the OS. Dedicated instance with 64 GB → around 54 GB for SQL.
sp_configure 'max server memory (MB)', <value>
TempDB with a single data file becomes a bottleneck under any concurrent workload. Rule: 1 data file per vCore up to 8, equal size and autogrowth across all files, TF 1117/1118 already default in SQL 2016+. A full TempDB is the most common incident we handle.
N data files = min(vCores, 8) · equal size · equal autogrowth
Query Store records plans, runtime stats and regressions — it is the black box of SQL Server 2016+. Without it enabled with QUERY_CAPTURE_MODE = AUTO, you rely on a volatile plan cache and guesswork. With it, plan regression becomes sp_query_store_force_plan in seconds.
ALTER DATABASE ... SET QUERY_STORE = ON (OPERATION_MODE = READ_WRITE)
Auto Update Statistics fires based on thresholds — on large tables those thresholds take too long to trigger and statistics get out of date. Traceflag 2371 (default in compat 130+) mitigates it, but scheduled UPDATE STATISTICS ... WITH FULLSCAN jobs are still non-negotiable.
UPDATE STATISTICS · sys.dm_db_stats_properties
Defaults MAXDOP = 0 and Cost Threshold = 5 are terrible on modern servers: any cheap query goes parallel, generating CXPACKET and high CPU. Tune according to the cores per NUMA node and raise Cost Threshold to 25–50.
sp_configure 'max degree of parallelism' · 'cost threshold for parallelism'
sys.dm_os_wait_stats shows where SQL Server spends its waiting time. PAGEIOLATCH = disk, SOS_SCHEDULER_YIELD = CPU, LCK_M_* = blocking, WRITELOG = log latency. Tuning without reading waits is guessing.
sys.dm_os_wait_stats · sys.dm_exec_session_wait_stats
SQL SERVER SECURITY HARDENING
SQL Server ships configured to work — not to be secure. Every default install is a list of vulnerabilities waiting for an audit.
Service accounts, applications and users holding sysadmin are the norm in environments that were never audited. Enforce least privilege: db_datareader, db_datawriter and custom roles with EXECUTE only on the procs actually needed.
SELECT * FROM sys.server_role_members WHERE role_principal_id = 3
SQL Server on public port 1433 is a favorite brute-force target. Restrict access at the firewall, use a non-default port on legacy instances, disable SQL Server Browser when named instances are not required and enforce Force Encryption = YES on the TCP/IP protocol.
Force Encryption = Yes · SQL Browser = Disabled
TDE (Transparent Data Encryption) protects data at rest — every mdf/ldf/backup is encrypted. Always Encrypted protects sensitive columns even from the DBA. TLS 1.2+ is mandatory for connections. Under GDPR/LGPD, missing encryption for personal data is clear evidence of negligence.
CREATE DATABASE ENCRYPTION KEY ... ALTER DATABASE ... SET ENCRYPTION ON
SQL Server Audit + Extended Events log logins, DDL, sensitive DML and access to protected objects. Without an audit trail there is no evidence of control for compliance and no forensic investigation possible after an incident.
CREATE SERVER AUDIT · CREATE DATABASE AUDIT SPECIFICATION
SQL SERVER HIGH AVAILABILITY
Synchronous Always On AG with an asynchronous replica in a distant region, quarterly failover tests, validated backups. That is the standard HTI deploys and sustains in financial customers.
Synchronous and asynchronous Availability Groups, with automatic failover and readable replicas to offload reads. Default choice for 99.99%+ SLAs on SQL Server 2016 Enterprise and later — also available on Standard (Basic AG, one DB per group).
FCI on top of Windows Server Failover Cluster with shared storage. Protects the whole instance (not just databases), preserves SQL Agent jobs, logins and credentials on failover. The right pick when the requirement is full-instance HA.
Replica in a geographically separate datacenter (async AG or Log Shipping), tested backups with documented RTO/RPO, numbered failover runbook. DR that has never been tested is not DR — it is hope.
RESULTS IN PRODUCTION
FINANCIAL · SQL SERVER · MISSION-CRITICAL
99.99%
SQL Server cluster operated by HTI with Always On Availability Groups, asynchronous replica on a remote site and DBSnoop monitoring. Failover tested quarterly. Zero transaction loss during real incidents.
RETAIL · SQL SERVER · BLACK FRIDAY
ZERODOWN
Focused peak preparation: tempdb tuning, MAXDOP review, critical index adjustments and Query Store in AUTO mode. SQL Server sustained the largest transactional volume in the customer's history with zero downtime.
INDUSTRY · SQL SERVER · CONSOLIDATION
40%
Consolidation of 22 instances into 6, edition reassessment (Enterprise → Standard where possible), analytical workloads migrated to PostgreSQL. Real licensing cut with no loss of performance or HA.
Our team has already resolved more than 596 production database incidents. 30-minute SLA. No prior contract required.
Trigger emergency support now →wa.me/551140636900 · +55 11 4063-6900 · 24/7
WHY HTI
HTI has consulted on SQL Server since its earliest versions in Brazilian corporate Windows environments. Our senior DBAs have decades of experience running ERPs, core banking, e-commerce and defense workloads — where an instance outage makes the news.
Our 24/7 NOC has already handled more than 596 production database incidents — high CPU, full tempdb, corruption, split-brain Always On, recurring deadlocks. Pattern recognition no Microsoft certification can deliver.
Every DBA operating production environments works from a restricted-access room at PIT, in São José dos Campos — 24/7 cameras, biometric access, dedicated rack. Security that goes beyond an internal policy.
Our proprietary behavioral monitoring tool goes beyond CPU and disk — it tracks ingestion patterns, access attempts to protected objects and anomalous SQL Server behavior. Built in-house, with no third-party tooling dependency.
FAQ
We support every version in active production use: SQL Server 2016, 2017, 2019 and 2022, plus Azure SQL Database, Azure SQL Managed Instance, SQL Server on VM (Azure and AWS EC2) and Amazon RDS for SQL Server. For legacy versions (2012, 2014) we run an assessment and plan the upgrade.
Remote DBA is the full outsourcing of SQL Server administration — HTI takes over the environment: 24/7 monitoring, incident response, proactive tuning, CU/GDR patch management, capacity planning and monthly reports. You get a senior SQL Server DBA team without the cost of full-time in-house hiring.
For emergencies, our SLA is 30 minutes in business hours (Mon–Fri, 9am–6pm BRT) and 2 hours outside that window, 24/7. Emergency support does not require a prior contract — reach us via WhatsApp or phone and it is billed by the hour.
The Health Check covers: performance (waits, tempdb, Query Store, missing/duplicate indexes, regressed plans), security (logins, permissions, TDE, network exposure, audit), HA/DR (Always On, backups, RTO/RPO), instance configuration (max memory, MAXDOP, Cost Threshold) and compliance. Deliverable: a report with a prioritized roadmap.
Consulting is a project with defined scope and timeline — health check, performance tuning, version migration, Always On rollout. Remote DBA is continuous operations — HTI runs the environment under a monthly SLA. Both can be contracted together or independently.
Yes. We support Azure SQL Database, Azure SQL Managed Instance, SQL Server on Azure VM, Amazon RDS for SQL Server and SQL Server on EC2. Being in the cloud does not remove the need for tuning, security and HA — the same problems exist in any environment, and licensing cost can be optimized with BYOL and right-sizing.
Yes. Migrating from SQL Server to PostgreSQL is a frequent path to reduce licensing. We run assessments with Microsoft DMA + proprietary tooling, plan T-SQL to PL/pgSQL conversion, regression testing and a controlled cutover. The goal is always to preserve performance and integrity — never migration for its own sake.
Yes, it is one of the most frequent incidents we handle. We resolve the root cause: file sizing and count, autogrowth configuration, TF 1117/1118, review of temp tables vs. table variables, spills from bad statistics and workloads generating excessive versioning.
Talk to a SQL Server Expert. Describe your environment — version, edition, size, biggest pain right now. In 30 minutes you know whether we are the right partner.
No commitment. Just technical clarity.
SQL Server versions & related technologies
GO DEEPER
Waits, plan cache, parameter sniffing, MAXDOP, Cost Threshold and how to hunt the guilty query with Query Store and DMVs.
Read article →MIGRATIONTechnical comparison between the three most-used enterprise relational engines — licensing, HA, replication and tooling.
Read article →MONGODBHow to properly size and monitor the WiredTiger cache to keep MongoDB responsive under production load.
Read article →