Can Flexprice Handle Pinecone's Pricing?

​

Pricing profile

​

Pinecone Standard database prices

​

What the units measure

• Reads: A query consumes 1 RU per GB of namespace size, with a minimum of 0.25 RU per query. Fetch consumes 1 RU per 10 records fetched, rounded up, with a minimum of 1 RU. A list request costs 1 RU. A request count alone cannot reproduce these charges.
• Writes: An upsert consumes 1 WU per KB of request data, with a minimum of 5 WUs per request. Replacing a record also counts the existing record’s size. Updates count new and existing record data; deletes count deleted record data, each with a 5 WU request minimum. Deleting an entire namespace or using deleteAll costs 5 WUs. Preserve Pinecone’s request-level rules instead of dividing all monthly bytes by a KB and applying one monthly minimum.
• Storage: An index’s size includes its records across all namespaces, including IDs, vector data and metadata. Storage has a duration component: Pinecone’s official worked example prorates storage for a four-hour workload. A namespace’s query size is not the organization’s total stored size.

​

How Pinecone’s pricing maps to Flexprice

​

Design takeaways

• Apply the floor after combining eligible charges. Pinecone’s minimum-usage examples show one shortfall line for the organization. Keep read, write and storage charges separate for explainability, but calculate their shared shortfall once. Independent $50 floors would introduce a $150 minimum that this offering does not have.
• Treat region as part of a price’s identity. The on-demand rate table varies by cloud and region. Filtering all four meters to the selected scope prevents usage from another region from silently receiving the AWS us-east-1 prices. A region field without a matching meter filter is only a label, not a pricing boundary.
• Normalize before using SUM. Rounded API read counts lose fractional consumption, while Pinecone’s storage example depends on elapsed time. SUM is appropriate here because the event carries reconciled RUs, WUs and GB-months, not rounded request responses or repeated storage snapshots. The normalization contract matters as much as the meter’s aggregation type.

​

Set up the monthly usage plan

​

1. Reconcile the billing quantities

1. Select only the Standard AWS us-east-1 on-demand database SKUs. Aggregate all indexes and projects in the organization once. Verify that excluded services have no charges and egress remains within the included 100 GB.
2. Obtain precise billed RU and WU quantities, retaining fractions and Pinecone’s operation-level minima. Do not use request counts or rounded response usage.
3. Normalize reconciled storage to GB-months. Respect the report’s actual billing unit and duration. If you only have the final pre-tax storage charge for this one SKU, dividing that charge by 0.33 gives its billed GB-month equivalent. This imports Pinecone’s storage calculation rather than inventing a sampling rule. The worked example below assumes a constant 100 GB for the entire month, equivalent to 100 GB-months.
4. Calculate the eligible usage subtotal from these three charges, then calculate max(50 - subtotal, 0) once for the organization. Emit one finalized monthly event after reconciling the report. Do not ingest Pinecone’s minimum line as usage and also calculate your own shortfall.

​

2. Create four metered features

​

3. Add the plan and usage charges

​

Send the reconciled usage and check the arithmetic

{
  "event_name": "pinecone_monthly_usage",
  "external_customer_id": "pinecone_org_example",
  "event_id": "pinecone_org_example_2026_10_final",
  "timestamp": "2026-10-31T23:59:59Z",
  "properties": {
    "pricing_scope": "standard_on_demand",
    "region": "aws_us_east_1",
    "read_units": 2000000,
    "write_units": 1000000,
    "storage_gb_months": 100,
    "minimum_shortfall_usd": 0
  }
}

from decimal import Decimal

D = Decimal
READ_RATE = D("16") / D("1000000")
WRITE_RATE = D("4") / D("1000000")
STORAGE_RATE = D("0.33")
MINIMUM = D("50")

def monthly_bill(read_units, write_units, storage_gb_months):
    subtotal = (
        D(read_units) * READ_RATE
        + D(write_units) * WRITE_RATE
        + D(storage_gb_months) * STORAGE_RATE
    )
    shortfall = max(MINIMUM - subtotal, D("0"))
    return subtotal, shortfall, subtotal + shortfall

assert monthly_bill("2000000", "1000000", "100") == (
    D("69"), D("0"), D("69")
)
assert monthly_bill("500000", "500000", "10") == (
    D("13.30"), D("36.70"), D("50.00")
)

​

Pricing sources

• Pinecone Database overview: database infrastructure for application retrieval
• Pinecone quickstart: application integration through API keys and developer tooling
• Upgrade your billing plan: the self-serve Standard console upgrade path used by this profile
• Pinecone pricing, Standard Database On-Demand regional table: AWS us-east-1 rates, monthly minimum and included egress
• Understand Pinecone cost: RU and WU definitions, storage composition and minimum usage examples
• Check usage per request: fractional consumption versus rounded API response counts
• Monitor performance: counter and gauge semantics and example scrape cadence
• Download a usage report: Standard console CSV, UTC dates and reporting delay
• Test Pinecone at scale: storage duration example

​

Related recipe