Compare

A Datadog Alternative for Oracle: No Agent on the Host, No SaaS

3 November 2026

Datadog Database Monitoring watches Oracle through the Datadog Agent, which its own architecture page says you "install directly onto the database host" for self-hosted databases, with a user, some forty grants and a view created as SYSDBA inside each database, and the data goes "through to Datadog's backend" over the internet. Scops is the most capable Oracle monitoring platform and goes the other way: nothing on the host, nothing created in the database, one read-only account through the listener, and every byte of data on your own server. Every session sampled every 5 seconds, 75+ Radar detectors, Scops AI with 40+ agentic tools and your own model, automation, Archiving and Compression, Standard Edition 2 in full, no Oracle pack, and a price per database rather than per host. Every Datadog statement below comes from Datadog's own documentation and price list.

At a glance

Criterion	Scops	Datadog Database Monitoring
On the database host	Nothing, ever. One read-only account with CREATE SESSION and SELECT_CATALOG_ROLE, through the listener.	The Datadog Agent 7.53.0 or later: "For self-hosted setups, you install the Agent directly onto the database host". For RAC, "configure the Agent for each RAC node".
Inside the database	Nothing.	A datadog user (c##datadog in a CDB), grants on some forty V$ and CDB views made as SYSDBA, and a view: "Log on as sysdba, create a new view in the sysdba schema, and give the Agent user access to it".
Where the data goes	Your server. Scops AI runs with your own model, so even the AI stays under your control.	"After connecting to the internet, it goes through to Datadog's backend." An on-premises option: not stated.
Detection	Radar Engine: 75+ detectors that forecast, catch pile-ups and anomalies, and name the cause. Each finding earns its level from the data.	Monitors you create on metrics, "out-of-the box monitors and SLO tracking", alerts for anomalies, outliers in a query's execution time.
Backups, jobs, accounts, objects	Built in: backups late or failing, scheduler jobs failing, passwords expiring, invalid objects and unusable indexes, NOLOGGING with a standby, the alert and listener logs.	Metrics for tablespaces, Data Guard apply lag and ASM disk groups; "blocking query insights". Backups, scheduler jobs, password expiry, invalid objects: not stated.
Session history	Every session every 5 seconds, on every database, kept; SQL statistics every 10 minutes; plan changes recorded.	"Query-level metrics, live and historical query snapshots, wait event analysis, database load, query explain plans". The collection interval and the retention are not stated.
AI	Scops AI, the most capable AI assistant for Oracle: 40+ agentic tools over the live database, the history and the findings, with your own model.	"An AI-powered root cause analysis agent" for query latency regressions and "automated query optimization recommendations". No assistant over live database state with your own model is described.
Automation	Automation, one-click fixes proposed by Scops AI, and the Archiving and Compression modules.	Recommendations with steps to remediate.
Oracle versions	12c to 26ai, Standard Edition 2 and Enterprise Edition, RAC and Data Guard, no Oracle pack.	"11g, 12c, 18c, 19c, 21c". 23ai and 26ai: not stated. Editions and packs: not stated; "running custom queries may result in additional costs or fees assessed by Oracle".
Price	Community: free for 2 databases. Professional: $499 per database a year, each further database 5% less.	$70 per database host a month on the list price page, billed annually: $840 a host a year, before the rest of Datadog.
Verdict	Everything a DBA watches, on your own server, with nothing on the hosts and nothing in the databases.	An Agent on every database host, a view in SYS, a SaaS backend, and a per-host bill.

What Datadog needs to run

From the Datadog Database Monitoring setup pages for self-hosted Oracle and the architecture page:

An Agent on each database host. "For self-hosted setups, you install the Agent directly onto the database host so that you have full visibility into the health of your system running the database process." For RAC, "customers must configure the Agent for each RAC node, because the Agent collects information from every node separately by querying V$ views". For a managed database such as Amazon RDS, the Agent goes on a separate host, and "it is recommended to have a single Datadog Agent monitor at most 30 database instances".
A user, forty grants and a view in every database. The setup page creates the datadog user, then, "Log on as sysdba", grants SELECT on V$SESSION, V$SQL, V$SQLSTATS, V$SQL_PLAN_STATISTICS_ALL, V$OSSTAT, V$DATAGUARD_STATS, V$ASM_DISKGROUP, DBA_OBJECTS, DBA_DATA_FILES and some thirty more, and then: "Log on as sysdba, create a new view in the sysdba schema, and give the Agent user access to it". In a container database the user is C##DATADOG with the SET CONTAINER privilege.
A direct path. "The Agent must connect directly to the host being monitored. The Agent should not connect to the database through a proxy, load balancer, or connection pooler."
The internet, and Datadog's backend. The architecture page's own description: the data "goes through to Datadog's backend". The intake endpoints are on your Datadog site. An on-premises option is not stated.
$70 per database host a month. Billed annually, on the list price page, for Database Monitoring alone.

Scops needs one Windows server with two services, used from a browser, and one read-only Oracle account per database with CREATE SESSION and SELECT_CATALOG_ROLE. Its repository is an Oracle schema on a database you already run, Enterprise Edition or Standard Edition 2. Nothing runs on the database hosts, nothing is created in the databases, nothing leaves your network, and the price is $499 per database a year after the two free ones.

What Datadog watches, in its own words

The integration page: "In addition to the standard integration features, Datadog DBM provides query-level metrics, live and historical query snapshots, wait event analysis, database load, query explain plans, and blocking query insights." The architecture page: "the Datadog Agent collects system metrics from the operating system host, database metrics directly from the database, and log events from database logs." Alerting is monitors: "create monitors to alert your team on Oracle database states", "out-of-the box monitors and SLO tracking", alerts for anomalies, and outliers "in a query's execution time or execution cost". The AI is "an AI-powered root cause analysis agent" for query latency regressions, with "automated query optimization recommendations". The metric list carries Oracle's tablespaces, Data Guard apply lag and ASM disk groups.

What the pages do not state: a collection interval, a retention period, an on-premises option, support for 23ai or 26ai, Oracle editions, whether any feature uses an Oracle pack, and any monitor for backups, scheduler jobs, password expiry or invalid objects. On custom queries the setup page warns that they "may result in additional costs or fees assessed by Oracle".

What Scops watches

Radar Engine, the detection layer of Scops, runs 75+ detectors on every database, and each one earns its level from the data rather than from a monitor you define:

Forecasts. Tablespace, temp, undo, ASM and recovery area space with the day they run out; datafiles against DB_FILES; CPU capacity; sequences near their MAXVALUE; partitions without a home.
Live pile-ups. Blocking chains to the root blocker, session surges, logon storms, slow single-block reads, slow commits, redo switches waiting for a checkpoint, waits that surge against the hour before, CPU saturation, the parallel pool, PGA at its limit, lost RAC blocks.
Anomalies. Each wait class and the CPU against the same hour on other days, plan changes tied to the statement they slowed, a statement taking most of the load, database growth far above the usual.
Operations. Backups late, failing or stuck; Data Guard transport and apply lag, apply stopped, NOLOGGING changes; scheduler and legacy jobs failing or overdue; passwords about to expire; invalid objects and unusable indexes; errors in the alert, ASM, Grid and listener logs.

On top of the detectors: every session every 5 seconds and SQL statistics every 10 minutes, kept on your server; Scops AI, the most capable AI assistant for Oracle, with 40+ agentic tools that read the live database, the history and the findings, explain why, propose the fix and change nothing without a click, running with your own model; automation for the routine work; and the Archiving and Compression modules that no monitoring tool has. The detector reference lists every detector with its thresholds.

Why Scops wins

Nothing on the host. No Agent to install on every database host and every RAC node, patch, and keep under one percent. One read-only account through the listener.
Nothing in the database. No user with forty grants, no SET CONTAINER, no view created as SYSDBA in each database. SELECT_CATALOG_ROLE, and that is all.
Your data stays yours. Sessions, SQL text, bind values and findings never leave your network, and the AI runs with your own model. Datadog's data goes to Datadog's backend over the internet.
Detection that names the cause. 75+ detectors that forecast, catch pile-ups and anomalies, with the cause and the level from the data, against monitors you define on metrics.
The rest of the DBA's day. Backups, jobs, accounts, schema objects, NOLOGGING and the logs are detectors in Scops; Datadog's pages state tablespaces, Data Guard lag, ASM and blocking, and nothing about the rest.
AI. The most capable AI assistant for Oracle, with 40+ tools over the live database and your own model, against a root-cause agent for query regressions.
Current Oracle, every edition. 12c to 26ai, Standard Edition 2 in full, no Oracle pack. Datadog lists 11g to 21c and states nothing about editions, packs or 26ai.
Price per database, not per host. Free for two databases, $499 a year per database after that, 5% less for each further one, against $840 a host a year for Database Monitoring alone.
Automation, Archiving and Compression. Routine work done for you, old rows moved online in rowid chunks, Oracle compression managed from one screen. Datadog's pages describe recommendations.
The interface. One window built for a DBA's day, Daily Overview first, instead of an observability suite in which Oracle is one database product of many.

Already running Datadog?

Keep it for the hosts, the containers and the applications; that is what it is for. Scops installs beside it in an afternoon for the Oracle side: a read-only account through the listener, with nothing on the hosts, so the Datadog Agent and its grants are untouched. Add the databases on day one, compare the two views for a fortnight, then drop the Database Monitoring line from the bill, the datadog user and the view from each database, and the per-node Agent configuration from the RAC nodes, on your own schedule.

Which one, in three lines

Datadog Database Monitoring: an Agent on every database host, a view in SYS, a SaaS backend, a per-host price, Oracle as one product of many.
Scops: the most capable Oracle monitoring platform, nothing on the hosts, nothing in the databases, your data on your server, a per-database price.
For Oracle, Scops. Datadog keeps the rest of the stack, and the Oracle side gets the detection, the AI and the price.

Questions

Is Scops a replacement for Datadog Database Monitoring on Oracle?
Yes, and more: the sessions, waits and SQL Datadog shows, sampled every 5 seconds, plus 75+ detectors, Scops AI, automation, Archiving and Compression, with no Agent on the host, nothing created in the database, and the data on your own server.

Does Datadog need an agent on the Oracle server?
For self-hosted Oracle, yes: Datadog's architecture page says to install the Agent directly onto the database host, and for RAC on each node. Scops has no agent; it connects through the listener with a read-only account.

Where does the monitoring data go?
With Datadog, to Datadog's backend over the internet; an on-premises option is not stated. With Scops, to an Oracle schema on your own server, and Scops AI runs with your own model.

What does Datadog Database Monitoring cost compared with Scops?
$70 per database host a month billed annually on Datadog's list price page, $840 a host a year, for Database Monitoring alone. Scops is free for two databases and $499 per database a year after that, with 5% off for each further database.