# The licence you are quoted is for the database. The cost you will carry is the migration — and nobody quotes that. *Databases & Data Tools — a TechBag decision guide. Last reviewed 2026-10-03.* > Engine licensing is the visible number, and it is usually the smaller one. The application rewrite, the driver differences, the stored procedures and the testing are where the project actually spends — and none of it appears on a vendor quote. **The checkable fact:** The test: ask whoever estimated the migration whether they counted the schema or the application. Schema conversion is largely automatable. Application behaviour is not, and that is the half that overruns. - Canonical: https://www.thetechbag.com/browse/devops/databases-data-tools - Category: [DevOps](https://www.thetechbag.com/browse/devops) - Products compared: 28 ## What databases & data tools actually is Twenty-eight products doing four distinct jobs. Fifteen **are** the database — an engine you license and run. Seven **monitor** one somebody else supplies. Five put schema changes **through the pipeline** the way application code already goes. One **models** the data for governance rather than operations. They are routinely conflated in a requirement, and the cost of conflating them is specific: buying monitoring when the problem was deployment friction, or buying an engine subscription when the actual gap was that nobody can see which query is slow. **The most common mis-purchase.** The **licence unit**. Per core, per instance, per developer or consumption — and **per core on virtualised infrastructure** is where estates get counted for hosts their database never used. Get the virtualisation counting policy in writing before signing anything per-core; it is the single most common surprise in this route. ## Four terms, resolved These are not tiers. A tool that monitors a database perfectly will not migrate it, and a migration path says nothing about whether the result is healthy. ### Database monitoring vs APM The query was slow, or the plan changed? ### Managed vs self-hosted vs DBaaS Where does the support boundary sit? ### Migration vs replication Different engine, or another copy of the same one? ### Compatibility vs conversion How much application code survives? The practical consequence: name the job before shortlisting. Half the products here cannot do the job the other half exists for, and the vocabulary hides it. ## The decision variables Seven variables move the shortlist. **Engine and version coverage.** Postgres, SQL Server, Oracle, MongoDB, MySQL — and what 'supported' means per tool, which varies from full diagnostics to a connection test. **The support boundary.** Self-managed, managed subscription on your infrastructure, or DBaaS. This decides the vendor list before any feature does. **Migration tooling and realistic effort.** Especially Oracle-to-Postgres, which is the live movement in Indian enterprises. Estimate from the application, not the schema. **Monitoring depth versus APM overlap.** Both name a slow query. Only one explains the plan. Decide whether you are buying both deliberately. **Licence model.** Per core, per instance, per developer; subscription or perpetual. Per core on virtualised hosts is the one to get in writing. **Compliance and audit reporting.** Whether the tool produces evidence an auditor accepts, or a screenshot somebody has to interpret. **India support presence and hours.** Undocumented across every vendor here. A support contract that does not cover Indian business hours is a real operational cost. ## The 28 products Every product carries one stated limitation. Prices are USD with an indicative INR conversion, tier- and term-matched where the vendor publishes one. ### Elasticsearch — Elastic - **Who it's for:** Search, log analytics and — increasingly — vector storage and retrieval for AI applications, on an engine a great many engineering teams already run. - **The honest limitation:** Operating it well at scale (shard strategy, index lifecycle, capacity, upgrades) is genuine engineering work. And AWS’s OpenSearch fork has diverged enough that documentation and clients are not interchangeable — check which one you deployed. - **Price:** Free → tier — free and open source under AGPL since 2024; paid subscription tiers add features and support. Elastic Cloud Hosted priced on provisioned resources across 60 regions, Serverless on consumption - **Details:** https://www.thetechbag.com/elastic/elasticsearch ### EDB Postgres Advanced Server — EDB - **Who it's for:** Enterprises with a live Oracle renewal and budget released to move off it — the most-proven compatibility path, and the one that keeps the most stored procedures intact. - **The honest limitation:** Compatibility reduces the rewrite; it does not remove it. The estimate that matters is the application's, not the schema's — and per-core licensing on virtualised infrastructure counts hosts in ways that surprise people. Confirm what the vendor counts. - **Price:** Per core, quote — quoted per core, subscription; Oracle-compatible Postgres — PL/SQL, Oracle syntax and package compatibility so a migration changes far less application code than plain Postgres would - **Details:** https://www.thetechbag.com/edb/edb-postgres-advanced-server ### EDB Postgres Extended Server — EDB - **Who it's for:** Estates already standardised on Postgres that want a supported enterprise distribution rather than community builds. - **The honest limitation:** No Oracle compatibility layer, so it is the wrong SKU for a migration project — that is Advanced Server. Buying this expecting PL/SQL support is the most common mix-up in the EDB line. - **Price:** Per core, quote — quoted per core; enterprise-hardened Postgres without the Oracle compatibility layer — for estates already on Postgres that need enterprise support, security and extended capability - **Details:** https://www.thetechbag.com/edb/edb-postgres-extended-server ### EDB Distributed High Availability — EDB - **Who it's for:** Estates whose Postgres cannot take a maintenance window — multi-site, active-active, with failover measured in seconds. - **The honest limitation:** Active-active is an architecture decision, not a switch: conflict handling and application behaviour must be designed for it. It also adds licensed cores rather than replacing them. - **Price:** Per core, quote — quoted per core as an addition to the Postgres subscription; active-active distributed replication for near-continuous availability across sites - **Details:** https://www.thetechbag.com/edb/edb-distributed-ha ### EDB Postgres AI — EDB - **Who it's for:** Postgres estates that want vector and analytical workloads beside the transactional data instead of a second platform to move it to. - **The honest limitation:** Newer than the core server line and the strongest case assumes EDB Postgres underneath. If the requirement is a dedicated vector database, a purpose-built one is deeper. - **Price:** Quote — quoted; the multi-model platform bringing analytics and AI workloads onto the same Postgres estate rather than exporting to a separate system - **Details:** https://www.thetechbag.com/edb/edb-postgres-ai ### MongoDB Atlas — MongoDB - **Who it's for:** Teams that want the document model without operating it — and who would rather the scaling decisions be a slider than a project. - **The honest limitation:** Consumption pricing is honest and hard to forecast: an inefficient query pattern shows up as a bill rather than a slow dashboard. Egress and cross-region transfer are billed separately. - **Price:** Consumption — consumption-priced per cluster on compute, storage and data transfer, with Indian cloud regions available; the vendor runs the database and the support boundary is the endpoint - **Details:** https://www.thetechbag.com/mongodb/mongodb-atlas ### MongoDB Database (self-managed) — MongoDB - **Who it's for:** Estates that must keep the data on their own infrastructure, for residency, latency or policy reasons. - **The honest limitation:** You own the upgrades, the backups, the sharding decisions and the on-call. The Atlas features arrive later, and some do not arrive at all. - **Price:** Quote — quoted for the Enterprise Advanced subscription; the same document database run on your own infrastructure, with the operational burden and the residency control both yours - **Details:** https://www.thetechbag.com/mongodb/mongodb-database ### MongoDB Atlas Vector Search — MongoDB - **Who it's for:** Teams building AI retrieval features that already store the source documents in MongoDB — one database, no synchronisation job. - **The honest limitation:** Tied to Atlas: the self-managed database is not the path here. A dedicated vector database will be deeper on index tuning and scale. - **Price:** In Atlas consumption — metered within Atlas consumption; vector storage and search beside the operational documents, so retrieval-augmented applications query one database rather than syncing two - **Details:** https://www.thetechbag.com/mongodb/mongodb-vector-search ### Redgate Monitor — Redgate - **Who it's for:** DBA teams responsible for many instances who need one place to see which of them is unhealthy right now. - **The honest limitation:** Instance-metered, so a sprawling estate of small databases costs like a large one. It watches the database and does not trace the application request that caused the load. - **Price:** Per instance, quote — quoted per monitored instance; estate-wide database health — waits, blocking, disk, backups and alerting across SQL Server, Postgres and Oracle from one console - **Details:** https://www.thetechbag.com/redgate/redgate-monitor ### Redgate Flyway — Redgate - **Who it's for:** Teams whose deployments stall because the schema change is a manual step performed by one person at the weekend. - **The honest limitation:** Per developer, so it scales with the team rather than the estate. It versions the schema; it does not tell you the production database is unhealthy. - **Price:** Per developer, quote — quoted per developer; schema migrations under version control so database changes ship through the same pipeline as application code, with a documented rollback path - **Details:** https://www.thetechbag.com/redgate/redgate-flyway ### Redgate Test Data Manager — Redgate - **Who it's for:** Regulated estates where a production restore into a test environment is the quiet compliance problem nobody has raised yet. - **The honest limitation:** Masking rules are yours to define and maintain, and they drift as the schema changes. It is a discipline with a tool, not a tool that supplies the discipline. - **Price:** Per developer, quote — quoted per developer; provisioning realistic test databases with sensitive fields masked, so non-production environments stop being a copy of production data - **Details:** https://www.thetechbag.com/redgate/redgate-test-data-manager ### Redgate SQL Toolbelt — Redgate - **Who it's for:** SQL Server shops whose developers work in SSMS all day and lose time to manual comparison and deployment scripting. - **The honest limitation:** SQL Server only. On a mixed estate it covers one engine and Flyway covers the rest, which is two licences for one job. - **Price:** Per developer, quote — quoted per developer; the SQL Server developer bundle — compare, data compare, prompt, source control and unit testing in one licence - **Details:** https://www.thetechbag.com/redgate/redgate-sql-toolbelt ### Idera SQL Diagnostic Manager — Idera - **Who it's for:** SQL Server estates that want deep engine-level diagnostics on their own infrastructure without a SaaS telemetry pipeline. - **The honest limitation:** SQL Server only and on-premises oriented. It is a specialist, so a mixed-engine estate needs a second tool beside it. - **Price:** Per instance, quote — quoted per monitored instance, perpetual or subscription; SQL Server performance monitoring with query-level diagnostics, blocking analysis and predictive alerting - **Details:** https://www.thetechbag.com/idera/idera-sql-diagnostic-manager ### Idera DB PowerStudio — Idera - **Who it's for:** DBA teams running a genuinely mixed estate who would rather learn one tool than four vendor consoles. - **The honest limitation:** Breadth over depth: it is not as deep on any single engine as that engine's specialist. Interface age shows against newer tools. - **Price:** Per developer, quote — quoted per developer; cross-platform DBA and development tooling — administration, development, tuning and space management across four engines from one interface - **Details:** https://www.thetechbag.com/idera/idera-db-powerstudio ### Idera ER/Studio — Idera - **Who it's for:** Organisations that need the data model documented and governed rather than inferred from whatever is in production. - **The honest limitation:** Modelling and governance, not operations: it will not tell you the database is slow. Its value depends on someone owning the models, which is a role rather than a licence. - **Price:** Per developer, quote — quoted per named user; enterprise data modelling — logical and physical models, lineage and a business glossary over a multi-engine estate - **Details:** https://www.thetechbag.com/idera/idera-er-studio ### Idera SQL Safe Backup — Idera - **Who it's for:** SQL Server estates whose backup window has stopped fitting the night, or whose restore time is the number that fails an audit. - **The honest limitation:** SQL Server only and focused on the database layer — see the Backup & Cyber Resilience guides for the estate-wide picture including immutability. - **Price:** Per instance, quote — quoted per instance; compressed, encrypted SQL Server backups with faster restores and policy-based scheduling across the estate - **Details:** https://www.thetechbag.com/idera/idera-sql-safe-backup ### Quest Foglight for Databases — Quest - **Who it's for:** Genuinely mixed estates that need one console across every engine rather than a specialist tool per database. - **The honest limitation:** Breadth means the per-engine depth is behind the single-engine specialists, and the console is dense. It is a platform decision, not a quick install. - **Price:** Per instance, quote — quoted per monitored instance; cross-platform database observability with workload analytics across the widest engine list here — SQL Server, Oracle, Postgres, MySQL, MongoDB and more - **Details:** https://www.thetechbag.com/quest/quest-foglight-databases ### Quest Foglight for SQL Server — Quest - **Who it's for:** SQL Server estates wanting Quest depth on one engine without the cross-platform console. - **The honest limitation:** Single engine, and on a mixed estate you end up buying the cross-platform edition anyway. Compare the two before signing. - **Price:** Per instance, quote — quoted per monitored instance; SQL Server-specific performance monitoring with wait-state analysis and change tracking on the engine - **Details:** https://www.thetechbag.com/quest/quest-foglight-sql-server ### Quest Foglight for Oracle — Quest - **Who it's for:** Oracle estates — including those planning an exit, where knowing which workloads are genuinely heavy shapes the migration order. - **The honest limitation:** Oracle only. If the plan is to leave Oracle, this is a tool for the journey rather than the destination. - **Price:** Per instance, quote — quoted per monitored instance; Oracle performance monitoring with wait-event analysis, RAC awareness and storage correlation - **Details:** https://www.thetechbag.com/quest/quest-foglight-oracle ### Quest Foglight Performance Investigator — Quest - **Who it's for:** Estates whose recurring incident is 'the same query got slow overnight' and nobody can prove what changed. - **The honest limitation:** An add-on rather than a standalone product, so it assumes Foglight underneath and adds to that bill. - **Price:** Per instance, add-on — quoted per instance as an addition to Foglight; query-level diagnostics with historical plan comparison — the layer that answers why the plan changed, not merely that it did - **Details:** https://www.thetechbag.com/quest/quest-foglight-performance-investigator ### SolarWinds Database Performance Analyzer — SolarWinds - **Who it's for:** DBA teams tuning SQL Server, Oracle, MySQL, PostgreSQL and more by wait time, agentless, from one self-hosted console. - **The honest limitation:** Runs on your own Windows or Linux server and you patch it; MongoDB is not supported; AI Query Assist sends masked queries out through Platform Connect. - **Price:** Quote — subscription per monitored instance, quoted — licences float between instances and can be shared with SolarWinds SQL Sentry; the site’s Database card starts at $142 per database a month without naming a product - **Details:** https://www.thetechbag.com/solarwinds/solarwinds-database-performance-analyzer ### IBM Db2 — IBM - **Who it's for:** Estates already running Db2 on IBM Z, Power or Linux (core banking, insurance, SAP) that want a supported upgrade path and a managed option. - **The honest limitation:** A smaller skills pool than Postgres or SQL Server in India, and the paid editions’ licence metric isn’t published, so get the counting rule for virtual cores in writing. - **Price:** Free → $630/mo — Db2 Community Edition free (4 cores, 16 GB); Db2 SaaS on IBM Cloud from $630 / month billed hourly, free 200 MB plan; Standard and Advanced editions quoted; Amazon RDS for Db2 - **Details:** https://www.thetechbag.com/ibm/ibm-db2 ### SAP HANA Cloud — SAP - **Who it's for:** SAP estates building analytics, extensions or AI agents on live SAP data, with one engine for transactional, vector and graph queries instead of a separate store per model. - **The honest limitation:** Its pull is strongest where SAP data already lives; outside an SAP landscape it is a proprietary engine with a smaller talent pool than Postgres or SQL Server. HANA Cloud availability in an Indian region is not documented, so confirm the region before signing. - **Price:** Quote (capacity units) — metered in SAP capacity units across compute, memory and storage; quoted by SAP or a partner, itemised by TechBag in INR with GST. The multi-model in-memory database (relational, vector, graph, spatial, JSON) SAP positions under Business Data Cloud - **Details:** https://www.thetechbag.com/sap/sap-hana-cloud ### Snowflake AI Data Cloud — Snowflake - **Who it's for:** Analytics and AI teams that want one managed platform for the warehouse, data engineering and AI agents, with Indian regions at US prices and no clusters to run. - **The honest limitation:** Consumption is easy to start and hard to forecast: an oversized or never-suspended warehouse shows up as a bill, and Cortex AI is metered separately in AI credits. There is no self-managed option and no Google Cloud region in India. - **Price:** $2.00 / credit (≈ ₹166) — per credit on demand in AWS Mumbai and Azure Central India (Pune): Standard $2, Enterprise $3, Business Critical $4, VPS $6, the same as US East; an XS warehouse burns 1 credit an hour, billed per second; storage $23 per TB a month - **Details:** https://www.thetechbag.com/snowflake/snowflake-ai-data-cloud ### Snowflake Postgres — Snowflake - **Who it's for:** Snowflake customers who want their transactional Postgres managed in the same account, governance and bill as their analytics, in an Indian region. - **The honest limitation:** Generally available only since February 2026 and tied to a Snowflake account, so it is newer and narrower than EDB or the hyperscalers’ managed Postgres; no Oracle compatibility layer, and storage costs about five times Snowflake’s analytic storage. - **Price:** Consumption — platform credits per instance-hour by size (high availability doubles it) at your edition’s credit price, plus storage at about $118 per TB a month in AWS Mumbai; Postgres 16 to 18, generally available since February 2026 - **Details:** https://www.thetechbag.com/snowflake/snowflake-postgres ### Databricks Lakehouse (Databricks SQL) — Databricks - **Who it's for:** Teams already doing data engineering or ML on Databricks that want SQL warehousing, AI/BI dashboards and Power BI on the same open, Unity Catalog-governed tables, in an Indian region on any of the three clouds. - **The honest limitation:** DBU usage is hard to forecast without test runs, serverless costs about 10% more in India than US East, and Classic and Pro add a separate cloud VM bill and minutes of start-up. No on-premises edition; Lakehouse//RT is still Beta. - **Price:** $0.70 / DBU (≈ ₹58) — SQL Serverless per DBU in US East, $0.77 in AWS Mumbai, Azure Central India and GCP Mumbai, cloud instances included; SQL Pro $0.55 and Classic $0.22 plus your own cloud VMs; billed per second after a one-minute minimum (October 2026 list) - **Details:** https://www.thetechbag.com/databricks/databricks-lakehouse ### Databricks Lakebase — Databricks - **Who it's for:** Databricks customers building apps or AI agents that need transactional Postgres beside lakehouse data, synced with Unity Catalog, with a branch per developer and computes that scale to zero. - **The honest limitation:** Only exists inside a Databricks workspace, and storage costs about three times plain managed Postgres in India. Synced tables bill pipeline DBUs, there is no Oracle compatibility layer, and Google Cloud has no Indian region yet. - **Price:** $0.092 / CU-hour — autoscaling compute per Capacity Unit-hour (2 GB RAM) on AWS Premium in US East, $0.104 in Mumbai and $0.126 on Azure Central India, 50% off until 31 January 2027; storage $0.375 per GB-month in India; Postgres 16 to 18, generally available since January 2026 - **Details:** https://www.thetechbag.com/databricks/databricks-lakebase ### Neon (from Databricks) — Databricks - **Who it's for:** Developers and startups wanting serverless Postgres 14 to 18 with a copy-on-write branch per change and idle computes that cost nothing, without a Databricks workspace or a cloud account of their own. - **The honest limitation:** No Indian region exists (eight AWS regions, Azure closed to new projects), so data sits abroad; check DPDP and sector rules first. SOC 2, HIPAA and an SLA come only on the Scale plan, and there is no INR price. - **Price:** Free → $0.106 / CU-hour — self-serve on neon.com: Free plan with 100 CU-hours per project, Launch $0.106 and Scale $0.222 per CU-hour (about 4 GB RAM each), storage $0.35 per GB-month, no monthly minimum; no Indian region, nearest AWS Singapore - **Details:** https://www.thetechbag.com/databricks/databricks-neon ## Why each constraint rules out what it does The full narrowing logic, as prose. Nothing here is hidden behind an interaction. **Postgres.** Rules out Elasticsearch, MongoDB Atlas, MongoDB Database (self-managed), MongoDB Atlas Vector Search, Redgate SQL Toolbelt, Idera SQL Diagnostic Manager, Idera SQL Safe Backup, Quest Foglight for SQL Server, Quest Foglight for Oracle, IBM Db2, SAP HANA Cloud, Snowflake AI Data Cloud and Databricks Lakehouse (Databricks SQL) — Postgres is not a covered engine. That leaves EDB Postgres Advanced Server, EDB Postgres Extended Server, EDB Distributed High Availability, EDB Postgres AI, Redgate Monitor, Redgate Flyway, Redgate Test Data Manager, Idera DB PowerStudio, Idera ER/Studio, Quest Foglight for Databases, Quest Foglight Performance Investigator, SolarWinds Database Performance Analyzer, Snowflake Postgres, Databricks Lakebase and Neon (from Databricks). **SQL Server.** Rules out Elasticsearch, EDB Postgres Advanced Server, EDB Postgres Extended Server, EDB Distributed High Availability, EDB Postgres AI, MongoDB Atlas, MongoDB Database (self-managed), MongoDB Atlas Vector Search, Quest Foglight for Oracle, IBM Db2, SAP HANA Cloud, Snowflake AI Data Cloud, Snowflake Postgres, Databricks Lakehouse (Databricks SQL), Databricks Lakebase and Neon (from Databricks) — SQL Server is not covered. That leaves Redgate Monitor, Redgate Flyway, Redgate Test Data Manager, Redgate SQL Toolbelt, Idera SQL Diagnostic Manager, Idera DB PowerStudio, Idera ER/Studio, Idera SQL Safe Backup, Quest Foglight for Databases, Quest Foglight for SQL Server, Quest Foglight Performance Investigator and SolarWinds Database Performance Analyzer. **Oracle.** Rules out Elasticsearch, EDB Postgres Extended Server, EDB Distributed High Availability, EDB Postgres AI, MongoDB Atlas, MongoDB Database (self-managed), MongoDB Atlas Vector Search, Redgate SQL Toolbelt, Idera SQL Diagnostic Manager, Idera SQL Safe Backup, Quest Foglight for SQL Server, IBM Db2, SAP HANA Cloud, Snowflake AI Data Cloud, Snowflake Postgres, Databricks Lakehouse (Databricks SQL), Databricks Lakebase and Neon (from Databricks) — Oracle is not covered. That leaves EDB Postgres Advanced Server, Redgate Monitor, Redgate Flyway, Redgate Test Data Manager, Idera DB PowerStudio, Idera ER/Studio, Quest Foglight for Databases, Quest Foglight for Oracle, Quest Foglight Performance Investigator and SolarWinds Database Performance Analyzer. **MongoDB.** Rules out Elasticsearch, EDB Postgres Advanced Server, EDB Postgres Extended Server, EDB Distributed High Availability, EDB Postgres AI, Redgate Monitor, Redgate Flyway, Redgate Test Data Manager, Redgate SQL Toolbelt, Idera SQL Diagnostic Manager, Idera DB PowerStudio, Idera SQL Safe Backup, Quest Foglight for SQL Server, Quest Foglight for Oracle, Quest Foglight Performance Investigator, SolarWinds Database Performance Analyzer, IBM Db2, SAP HANA Cloud, Snowflake AI Data Cloud, Snowflake Postgres, Databricks Lakehouse (Databricks SQL), Databricks Lakebase and Neon (from Databricks) — MongoDB is not covered. That leaves MongoDB Atlas, MongoDB Database (self-managed), MongoDB Atlas Vector Search, Idera ER/Studio and Quest Foglight for Databases. **the database itself.** Rules out Redgate Monitor, Redgate Flyway, Redgate Test Data Manager, Redgate SQL Toolbelt, Idera SQL Diagnostic Manager, Idera DB PowerStudio, Idera ER/Studio, Idera SQL Safe Backup, Quest Foglight for Databases, Quest Foglight for SQL Server, Quest Foglight for Oracle, Quest Foglight Performance Investigator and SolarWinds Database Performance Analyzer — a tool that works on a database somebody else supplies. That leaves Elasticsearch, EDB Postgres Advanced Server, EDB Postgres Extended Server, EDB Distributed High Availability, EDB Postgres AI, MongoDB Atlas, MongoDB Database (self-managed), MongoDB Atlas Vector Search, IBM Db2, SAP HANA Cloud, Snowflake AI Data Cloud, Snowflake Postgres, Databricks Lakehouse (Databricks SQL), Databricks Lakebase and Neon (from Databricks). **database monitoring.** Rules out Elasticsearch, EDB Postgres Advanced Server, EDB Postgres Extended Server, EDB Distributed High Availability, EDB Postgres AI, MongoDB Atlas, MongoDB Database (self-managed), MongoDB Atlas Vector Search, Redgate Flyway, Redgate Test Data Manager, Redgate SQL Toolbelt, Idera DB PowerStudio, Idera ER/Studio, Idera SQL Safe Backup, IBM Db2, SAP HANA Cloud, Snowflake AI Data Cloud, Snowflake Postgres, Databricks Lakehouse (Databricks SQL), Databricks Lakebase and Neon (from Databricks) — not a monitoring product. That leaves Redgate Monitor, Idera SQL Diagnostic Manager, Quest Foglight for Databases, Quest Foglight for SQL Server, Quest Foglight for Oracle, Quest Foglight Performance Investigator and SolarWinds Database Performance Analyzer. **database DevOps.** Rules out Elasticsearch, EDB Postgres Advanced Server, EDB Postgres Extended Server, EDB Distributed High Availability, EDB Postgres AI, MongoDB Atlas, MongoDB Database (self-managed), MongoDB Atlas Vector Search, Redgate Monitor, Idera SQL Diagnostic Manager, Idera ER/Studio, Quest Foglight for Databases, Quest Foglight for SQL Server, Quest Foglight for Oracle, Quest Foglight Performance Investigator, SolarWinds Database Performance Analyzer, IBM Db2, SAP HANA Cloud, Snowflake AI Data Cloud, Snowflake Postgres, Databricks Lakehouse (Databricks SQL), Databricks Lakebase and Neon (from Databricks) — not database DevOps tooling. That leaves Redgate Flyway, Redgate Test Data Manager, Redgate SQL Toolbelt, Idera DB PowerStudio and Idera SQL Safe Backup. **migration tooling.** Rules out Elasticsearch, EDB Postgres Extended Server, EDB Postgres AI, MongoDB Atlas, MongoDB Database (self-managed), Redgate Flyway, Idera DB PowerStudio, Idera ER/Studio, IBM Db2, SAP HANA Cloud, Snowflake AI Data Cloud, Snowflake Postgres, Databricks Lakehouse (Databricks SQL), Databricks Lakebase and Neon (from Databricks) — some import tooling, but no documented engine-migration path; EDB Distributed High Availability, MongoDB Atlas Vector Search, Redgate Monitor, Redgate Test Data Manager, Redgate SQL Toolbelt, Idera SQL Diagnostic Manager, Idera SQL Safe Backup, Quest Foglight for Databases, Quest Foglight for SQL Server, Quest Foglight for Oracle, Quest Foglight Performance Investigator and SolarWinds Database Performance Analyzer — no migration tooling. That leaves EDB Postgres Advanced Server. **a managed service.** Rules out EDB Distributed High Availability, MongoDB Database (self-managed), Redgate Monitor, Redgate Flyway, Redgate Test Data Manager, Redgate SQL Toolbelt, Idera SQL Diagnostic Manager, Idera DB PowerStudio, Idera ER/Studio, Idera SQL Safe Backup, Quest Foglight for Databases, Quest Foglight for SQL Server, Quest Foglight for Oracle, Quest Foglight Performance Investigator and SolarWinds Database Performance Analyzer — no managed service: the operational burden is yours. That leaves Elasticsearch, EDB Postgres Advanced Server, EDB Postgres Extended Server, EDB Postgres AI, MongoDB Atlas, MongoDB Atlas Vector Search, IBM Db2, SAP HANA Cloud, Snowflake AI Data Cloud, Snowflake Postgres, Databricks Lakehouse (Databricks SQL), Databricks Lakebase and Neon (from Databricks). **on-premises.** Rules out Elasticsearch, MongoDB Atlas, MongoDB Atlas Vector Search, SolarWinds Database Performance Analyzer, SAP HANA Cloud, Snowflake AI Data Cloud, Snowflake Postgres, Databricks Lakehouse (Databricks SQL), Databricks Lakebase and Neon (from Databricks) — managed service only. That leaves EDB Postgres Advanced Server, EDB Postgres Extended Server, EDB Distributed High Availability, EDB Postgres AI, MongoDB Database (self-managed), Redgate Monitor, Redgate Flyway, Redgate Test Data Manager, Redgate SQL Toolbelt, Idera SQL Diagnostic Manager, Idera DB PowerStudio, Idera ER/Studio, Idera SQL Safe Backup, Quest Foglight for Databases, Quest Foglight for SQL Server, Quest Foglight for Oracle, Quest Foglight Performance Investigator and IBM Db2. **not per-core.** Rules out EDB Postgres Advanced Server, EDB Postgres Extended Server, EDB Distributed High Availability and IBM Db2 — per-core licensing, which on virtualised infrastructure counts hosts in ways that surprise people. That leaves Elasticsearch, EDB Postgres AI, MongoDB Atlas, MongoDB Database (self-managed), MongoDB Atlas Vector Search, Redgate Monitor, Redgate Flyway, Redgate Test Data Manager, Redgate SQL Toolbelt, Idera SQL Diagnostic Manager, Idera DB PowerStudio, Idera ER/Studio, Idera SQL Safe Backup, Quest Foglight for Databases, Quest Foglight for SQL Server, Quest Foglight for Oracle, Quest Foglight Performance Investigator, SolarWinds Database Performance Analyzer, SAP HANA Cloud, Snowflake AI Data Cloud, Snowflake Postgres, Databricks Lakehouse (Databricks SQL), Databricks Lakebase and Neon (from Databricks). **an Indian region.** Rules nothing out on published terms. It flags EDB Postgres Advanced Server — An Indian region or in-country support hours are not documented on the vendor's pages, EDB Postgres Extended Server — An Indian region or in-country support hours are not documented on the vendor's pages, EDB Distributed High Availability — An Indian region or in-country support hours are not documented on the vendor's pages, EDB Postgres AI — An Indian region or in-country support hours are not documented on the vendor's pages, MongoDB Database (self-managed) — An Indian region or in-country support hours are not documented on the vendor's pages, Redgate Monitor — An Indian region or in-country support hours are not documented on the vendor's pages, Redgate Flyway — An Indian region or in-country support hours are not documented on the vendor's pages, Redgate Test Data Manager — An Indian region or in-country support hours are not documented on the vendor's pages, Redgate SQL Toolbelt — An Indian region or in-country support hours are not documented on the vendor's pages, Idera SQL Diagnostic Manager — An Indian region or in-country support hours are not documented on the vendor's pages, Idera DB PowerStudio — An Indian region or in-country support hours are not documented on the vendor's pages, Idera ER/Studio — An Indian region or in-country support hours are not documented on the vendor's pages, Idera SQL Safe Backup — An Indian region or in-country support hours are not documented on the vendor's pages, Quest Foglight for Databases — An Indian region or in-country support hours are not documented on the vendor's pages, Quest Foglight for SQL Server — An Indian region or in-country support hours are not documented on the vendor's pages, Quest Foglight for Oracle — An Indian region or in-country support hours are not documented on the vendor's pages, Quest Foglight Performance Investigator — An Indian region or in-country support hours are not documented on the vendor's pages, SolarWinds Database Performance Analyzer — An Indian region or in-country support hours are not documented on the vendor's pages, IBM Db2 — An Indian region or in-country support hours are not documented on the vendor's pages, SAP HANA Cloud — An Indian region or in-country support hours are not documented on the vendor's pages and Neon (from Databricks) — An Indian region or in-country support hours are not documented on the vendor's pages — marked, not removed. **Per core is the licensing trap.** On virtualised infrastructure, per-core counting frequently includes cores your database never uses — the host's, not the VM's. Confirm in writing what the vendor counts before signing anything per-core, and get the virtualisation policy in the contract. **Migration is estimated from the wrong artefact.** Teams estimate an Oracle-to-Postgres move from the schema, because the schema is easy to count. The cost lives in the application: stored procedures, driver behaviour, SQL dialect, transaction assumptions. EDB's compatibility layer reduces that work materially; it does not remove it. **Monitoring gets bought twice.** APM names the slow query from the outside; database monitoring explains it from inside the engine. Estates commonly own both without deciding to. That is defensible — but decide it, rather than discovering it at renewal. **Open source is not free at the support boundary.** Postgres, MySQL and MongoDB community editions cost nothing to download. The number that appears later is the support contract, and it appears at the worst possible moment. Budget it at the start. ## Eight situations, and what each one buys If one of these is your sentence, the shortlist is short. ### An Oracle renewal arrived and somebody said Postgres **Shortlist:** EDB Postgres Advanced Server, Quest Foglight for Oracle **Why:** Advanced Server is the most-proven compatibility path; Foglight for Oracle tells you which workloads are heavy, which decides the migration order. **Trade-off:** Compatibility reduces the rewrite rather than removing it — read the depth section below before committing to a date. ### We run four different engines and four different consoles **Shortlist:** Quest Foglight for Databases, Idera DB PowerStudio **Why:** Both cover a genuinely mixed estate from one place — Foglight for monitoring, PowerStudio for administration. **Trade-off:** Breadth trades against per-engine depth; the specialists see more on any single engine. ### Schema changes are a manual step someone does at the weekend **Shortlist:** Redgate Flyway, Redgate SQL Toolbelt **Why:** Migrations under version control, shipping through the same pipeline as application code, with a documented rollback. **Trade-off:** Per developer, so it scales with team size. Toolbelt is SQL Server only; Flyway spans engines. ### We do not want to run a database at all **Shortlist:** MongoDB Atlas, EDB Postgres AI **Why:** The vendor owns upgrades, backups and scaling; you get an endpoint and a support boundary. **Trade-off:** Consumption pricing is honest and hard to forecast — an inefficient query becomes a bill rather than a slow dashboard. ### The data cannot leave our infrastructure **Shortlist:** MongoDB Database (self-managed), EDB Postgres Advanced Server, Idera SQL Diagnostic Manager **Why:** All three run entirely on your own infrastructure with a commercial support contract behind them. **Trade-off:** You own the upgrades and the on-call — the operational burden the managed price was covering. ### The same query got slow overnight and nothing changed **Shortlist:** Quest Foglight Performance Investigator, Idera SQL Diagnostic Manager, Redgate Monitor **Why:** Historical plan comparison is the only thing that proves what changed — usually statistics or an index, not the code. **Trade-off:** Performance Investigator is an add-on and assumes Foglight underneath. ### Our test environment is a copy of production, including the personal data **Shortlist:** Redgate Test Data Manager **Why:** Provisioning realistic test databases with sensitive fields masked, which is the quiet compliance problem in most estates. **Trade-off:** Masking rules are yours to define and they drift as the schema changes. ### We are building retrieval features on our own documents **Shortlist:** MongoDB Atlas Vector Search, EDB Postgres AI **Why:** Vectors beside the operational data means one database and no synchronisation job between two systems. **Trade-off:** Both assume their platform underneath, and a dedicated vector database will be deeper on index tuning at scale. ## At scale Database tooling scales by instance count and engine variety, not by user count. ### 1 database instances — A handful of instances, one engine - Native tooling plus a specialist is enough - Per-instance pricing is affordable at this size - Schema changes can still be manual, just about **The test:** Version the schema before the team grows, not after. ### 2 database instances — Dozens of instances, one or two engines - Estate-wide monitoring starts to pay - Per-instance costs become visible - Test data and masking become a real question **The test:** Count instances honestly — sprawl is what moves this bill. ### 3 database instances — Mixed engines at scale - Cross-platform consoles beat four specialists - Licence models differ per engine and complicate renewal - Modelling and lineage start to matter **The test:** Compare the cross-platform edition against the specialists you already own. ### 4 database instances — Enterprise, with a migration in flight - Per-core counting on virtualised hosts becomes material - Monitoring both old and new engines through the transition - Support hours and escalation paths become contractual **The test:** Get the virtualisation counting policy in the contract, in writing. Where a vendor does not publish list pricing, this page says so rather than implying a figure. ## Getting out The database is the hardest thing on this site to leave, and the tooling around it is not. **The data** — Every engine exports; the format and the downtime are the negotiation *(Portable, with a window)* **Stored procedures and dialect SQL** — Rewritten unless the target offers a compatibility layer *(The real cost)* **Monitoring configuration** — Thresholds, baselines and alert routing are re-authored per tool *(Rebuilt)* **Schema version history** — Flyway and similar keep migrations in your repository, not the vendor's *(Portable if versioned)* The practical consequence: version your schema migrations in your own repository from day one. It is the one artefact here that stays yours whatever you buy next. ## What it costs By licence unit, in USD and INR, with the counting policy named. ### Do you already own one? Four checks, in the order most likely to return a yes. - **Your cloud provider — RDS, Azure SQL and Cloud SQL include basic monitoring.** Enough for health and simple alerting; they **stop** at wait-state and plan-level diagnosis. - **Native engine tooling — Query Store, pg_stat_statements and AWR are already there.** Genuinely useful and free, and they need somebody who knows how to read them. - **Your APM — It already names the slow query.** Different depth, same symptom. See the boundary section before buying database monitoring as well. - **Community editions — Postgres, MySQL and MongoDB cost nothing to download.** The support contract is the number that appears later. Budget it at the start, not at the incident. Native tooling is stronger in this category than most buyers assume. The case for a commercial tool is usually estate-wide visibility, not depth on one instance. ### What the rest actually cost Four licence units, and where each one surprises people. ### What isn't in the licence price - **Migration effort.** The largest line in any engine change, and it never appears on the vendor quote. Estimate from the application. - **Support contracts on open source.** Community editions are free to download. Production support is not, and it is needed at the worst moment. - **Per-core counting on virtual hosts.** Cores your database never used, counted because the hypervisor could have scheduled them there. - **The DBA who reads the output.** Every monitoring tool here produces more signal than an unowned console will ever action. ## What goes wrong Five ways this purchase goes wrong. - **Per-core licensing counted against the host, not the VM.** The estate is billed for cores the database never used. Get the virtualisation counting policy in the contract before signing. - **Migration estimated from the schema.** Tables convert; PL/SQL, driver behaviour and application assumptions do not. The schema is the easy half and the cheap half. - **Monitoring bought twice.** Once inside APM and once at the database layer. Defensible if deliberate — expensive when discovered at renewal. - **Support hours that miss Indian business hours.** A severity-one at 10am IST answered at 9am Pacific is a working day lost. Confirm coverage, not just the SLA number. - **Open source assumed free, then the support contract arrives.** Usually mid-incident, at a price nobody budgeted. Decide the support position before production, not during it. ## Questions this guide answers ### What does an Oracle-to-Postgres migration actually cost? The licence saving is real and survives scrutiny; the migration cost is where estimates fail. Schema conversion — tables, indexes, constraints — is largely automatable. What is not: PL/SQL packages, Oracle-specific SQL, driver behaviour, sequence and transaction semantics, and every application that assumed Oracle. EDB Postgres Advanced Server accepts PL/SQL and Oracle syntax, which materially reduces that second half; it does not remove it. Estimate from the application, not the schema, and run an assessment before committing to a date. ### What is the difference between database monitoring and APM? APM traces the request from outside and names the slow query. Database monitoring sits inside the engine: wait states, execution plans, index health, blocking chains. APM rarely tells you the statistics went stale or the plan flipped under a parameter sniff; database monitoring cannot follow a request across six services. Many estates own both, which is defensible — the question is whether they chose to or discovered it. ### How are database tools licensed? Four different units on this page. EDB quotes per core by subscription. Redgate quotes per developer for Flyway, Test Data Manager and SQL Toolbelt, and per instance for Monitor. Idera and Quest quote per monitored instance. MongoDB Atlas is consumption-priced. Per core is the one that surprises people: on virtualised infrastructure the vendor may count the host's cores rather than the VM's, so get the counting policy in writing. ### Is community Postgres good enough without a commercial subscription? For many workloads, yes — Postgres is genuinely production-grade. What you are buying with a commercial subscription is a support boundary and, in EDB Advanced Server's case, Oracle compatibility. The mistake is assuming free means free: the support contract is needed at the worst possible moment, and negotiating it during an incident costs more than budgeting it at the start. ### Do these tools support Indian business hours? It is not documented on the vendor pages for any product on this page, which is why we mark it unknown rather than assume either way. It is a real operational cost — a severity-one raised at 10am IST and answered at 9am Pacific is a working day lost. Ask for coverage hours and the escalation path in the contract, not just the SLA number, and TechBag will confirm it as part of the quote. --- *Vendor-neutral. Nothing gated. Figures marked [TechBag to confirm] are team-owned and deliberately not guessed.* *TechBag — India-based software reseller and advisory. https://www.thetechbag.com/browse/devops/databases-data-tools*