datadog-org-cost-report

Datadog Org Cost Report

Estimate Datadog list-price usage cost for any group of internal orgs, queried from Snowflake. Encodes the workflow and query patterns used to produce the February 2026 LMS org cost analysis (~$775K for ~7,567 active orgs).

---

Inputs

Collect two pieces of information before running queries:

1. Org group identifier (one of two mutually exclusive options):

• parent_org_id (preferred): filter by DIM_ORG.MASTER_ORG_ID (e.g., 17166 for LMS/curriculum orgs). More reliable than name patterns: covers renamed orgs, avoids false matches on overlapping suffixes.
• org_name_pattern (fallback): SQL LIKE pattern applied to DIM_ORG.NAME (e.g., %-LMS, %-DEMO, %-PARTNER). Use when the parent org ID is unknown.

2. month (optional): target billing month in YYYY-MM format. Defaults to the most recent complete calendar month (UTC).

Known parent org IDs:

• 17166 = LMS/curriculum orgs (the org provisioned by the training platform)

---

Scope CTE Pattern

Every query in a session must open with the same org_scope CTE so that the org group definition is consistent across all queries. Two variants — pick one and use it throughout the session:

Variant A: by parent org ID (preferred)

WITH org_scope AS (
  SELECT DISTINCT ID AS ORG_ID, NAME, DATACENTER
  FROM REPORTING.GENERAL.DIM_ORG
  WHERE MASTER_ORG_ID = 17166   -- replace with the actual parent org ID
)

Variant B: by name pattern (fallback)

WITH org_scope AS (
  SELECT DISTINCT ID AS ORG_ID, NAME, DATACENTER
  FROM REPORTING.GENERAL.DIM_ORG
  WHERE NAME LIKE '%-LMS'       -- replace with the actual LIKE pattern
)

Ready-to-run SQL for all workflow phases is in references/queries.md.

---

Core Principles

1. Trial/free plans are not supported by the GA cost model. Most internal orgs are on trial or free plans. BYOD_COST_GA_V2_US1_MICHELADA returns "unsupported" for these orgs. Use list prices from DIM_STARTUP_PRICING_BY_PRODUCT_AGGREGATION for all cost estimates instead.

2. Host costs use monthly average as a lower-bound proxy. FACT_ORG_ALL_PRODUCT_USAGE_MONTHLY stores cumulative hourly host snapshots (host_count_monthly_sum). Datadog's actual billing uses the 99th percentile. Dividing the monthly sum by 744 (hours in a 31-day month) yields an average, which is a conservative lower bound.

3. Estimates are list prices only. Enterprise customers receive significant discounts. The list-price total represents "retail value," not invoice cost.

4. Not all products have prices in the startup pricing table. SIEM, Flex Logs, and Serverless are excluded from cost estimates (no matching unit price). Include their raw usage quantities in the report but mark them as "no price available."

---

Quick Start

Step 1: Load Snowflake MCP Use ToolSearch to load the Snowflake MCP server before running any queries.

Step 2: Build org_scope; get active org count + plan type breakdown Use Query 2 from references/queries.md. Confirms scope and flags plan distribution.

Step 3: Coverage check Use Query 3 to see which billing dimensions are present in the data and which are excluded (no unit price). Sets expectations before computing costs.

Step 4: Query monthly usage and apply unit prices Use Query 4 (monthly cost summary) from references/queries.md.

Step 5: Write report Write results to reports/org-cost-{label}-{YYYY-MM}.md relative to the project root. Use {label} as a short identifier for the org group (e.g., lms, demo). Include the mandatory caveats block (see below).

---

Mandatory Report Caveats

Every generated report must include this block (adapt values as needed):

## Caveats

* **List prices only.** Enterprise customers receive significant discounts.
  This estimate reflects retail/list pricing, not actual invoice cost.
* **Average vs 99th-percentile hosts.** Host costs are estimated using monthly
  average (sum / 744 hours). Datadog bills on 99th percentile, which is higher.
  These figures are a lower bound.
* **Unsupported products excluded.** SIEM, Flex Logs, and Serverless do not
  have unit prices in the startup pricing table and are excluded from cost totals.
* **US1 scope.** `BYOD_COST_GA_V2_US1_MICHELADA` covers US1 only. Usage from
  EU1, AP1, US3, and US5 orgs may be incomplete or missing.
* **Pricing source:** `REPORTING.GENERAL.DIM_STARTUP_PRICING_BY_PRODUCT_AGGREGATION`
* **Usage source:** `REPORTING.GENERAL.FACT_ORG_ALL_PRODUCT_USAGE_MONTHLY`

---

Key Tables

Table	Schema	Purpose
DIM_ORG	REPORTING.GENERAL	Org metadata: ID, NAME, MASTER_ORG_ID, DATACENTER
FACT_ORG_ALL_PRODUCT_USAGE_MONTHLY	REPORTING.GENERAL	Monthly usage by org and billing dimension; join on ORG_ID = DIM_ORG.ID
DIM_STARTUP_PRICING_BY_PRODUCT_AGGREGATION	REPORTING.GENERAL	Unit prices by PRODUCT_AGGREGATION; join on PRODUCT_AGGREGATION = PRIMARY_BILLING_DIMENSION
BYOD_COST_GA_V2_US1_MICHELADA	REPORTING.GENERAL	GA V2 cost model; returns "unsupported" for trial/free orgs -- use for plan type confirmation only

---

Report Output Format

# Datadog Org Cost Report: {Label} Orgs - {Month}

## Summary
| Metric | Value |
|---|---|
| Total orgs in scope | X |
| Active orgs (usage > 0) | X |
| Total estimated cost (list price) | $X |
| Cost per active org | $X |

## Cost Breakdown by Product
| Product | Metric | Usage | Unit Price | Est. Cost |
|---|---|---|---|---|
| Infrastructure Monitoring | avg hosts | X | $18/host | $X |
| APM | avg hosts | X | $36/host | $X |
| ... | ... | ... | ... | ... |
| **Total** | | | | **$X** |

## Products Without Pricing
| Product | Usage | Note |
|---|---|---|
| SIEM | X events | No unit price available |
| Flex Logs | X events | No unit price available |
| Serverless | X invocations | No unit price available |

## Caveats
[mandatory caveats block]

## Metabase Queries
[inline SQL for 6-month trend, per-org breakdown, product mix]

---

References

• references/pricing.md -- static unit price table from DIM_STARTUP_PRICING_BY_PRODUCT_AGGREGATION
• references/queries.md -- ready-to-run SQL templates for all workflow phases
• references/definitions.md -- canonical metric definitions (active org, month boundary, host proxy, etc.)