Run This Observability Vendor Check Before Standardizing on Datadog

If you are about to standardize on an observability platform, treat it like a company dependency. The dashboards are the easy part; the collection layer, billing model, and data location decide whether the stack survives your next hiring round, incident, and procurement review.

A large vendor is a useful worked example because its seams are visible. Datadog can be an excellent platform and still become a hard dependency if the checks below are skipped.

The five checks that decide whether standardization is safe

Use this list before standardization, ordered by the way problems usually surface: collection, billing, data location, log coverage, and enterprise readiness.

• Ask what breaks if you remove the vendor agent.
• Model a high-cardinality month before you sign.
• Put telemetry in your own bucket and see what the vendor still requires.
• Search logs outside the platform and confirm the query path is real.
• Check enterprise and compliance signals that matter to your customers.

Agent coverage defines portability

The collection layer is the part of the stack you cannot see during a demo. When the features you actually use require the vendor agent, you have less leverage than the dashboard suggests.

Before standardizing, map which dashboards, alerts, and integrations depend on the agent. If the answer includes your core incident workflow, treat the agent as infrastructure, not a plugin.

Database monitoring at an advanced level, data observability, and cloud network monitoring in Datadog require its own agent or its own OpenTelemetry distribution.

Cost, storage, and log coverage are the exit levers

Billing surprises tend to follow workload changes. Datadog's infrastructure monitoring runs $15 to $23 per host per month on annual plans, then adds charges for custom metrics, ingested logs, and indexed traces.

Build a test workload with the tags you actually plan to use, then compare the bill against a simpler baseline. The same workload can generate very different bills once custom metrics and high-cardinality tags are in use.

Cost pressure is not hypothetical. A survey of 500 U.S. technology professionals in April 2026 found that 49% of technology leaders put AI workloads at 26% to 50% of total observability costs, 42% reported overages, and 79% had adopted at least one cost-control step.

Data location defines your exit. Datadog now offers a bring-your-own-cloud option that keeps metrics, logs, and traces in the customer's cloud while Datadog handles analysis.

Ask where the raw telemetry lives before the vendor's UI sees it. When it leaves your account, you need to understand retention, access, and deletion. When it stays in your account, you still need to know what the vendor can query and what it cannot.

In a BYOC setup, advanced features still need Datadog's agent or its version of the OpenTelemetry collector, while upstream OpenTelemetry covers only baseline functionality.

With BYOC, raw telemetry can remain in customer cloud object storage, where storage costs are about $0.023 per gigabyte per month and the customer pays that rate.

Logs often already live elsewhere. Datadog's Log Explorer now lets users search logs across Databricks, ClickHouse, and Snowflake.

This matters when your data platform already owns the logs. You want the observability tool to query the store, not create another copy with its own retention rules.

Enterprise signals show whether the vendor can stay

Regulated customers need a vendor that can remain in their environment. On May 27, 2026, Datadog for Government received FedRAMP High certification covering its observability and security platform.

Compliance is a baseline, not a complete answer. Check whether the certification covers the services your customers will actually use.

Scale matters too, because it tells you whether the vendor has the resources to keep building. Datadog's fiscal 2026 first-quarter revenue grew 32% year over year to $1.01 billion, while its number of customers with annual recurring revenue over $100,000 rose 21% to 4,550.

Datadog's fourth-quarter 2025 revenue reached $953 million, up 29% year over year, and its customer count with annual recurring revenue above $1 million rose to 603 from 462 a year earlier.

More than 6,500 Datadog customers used at least one AI integration, and that group accounted for about 80% of annual recurring revenue.

Goldman Sachs cut Datadog from Buy to Sell on January 12, 2026, with a $113 price target, citing AI-driven telemetry volumes, competitors focused on reducing observability budgets, and Amazon's cloud observability push.

If the answers are unclear, keep the platform in a pilot and revisit the same questions after a full billing cycle.