Google Cloud Healthcare · Rate Limits

Google Cloud Healthcare Rate Limits

Cloud Healthcare API enforces quotas measured per minute per region across FHIR, HL7v2, and DICOM store types. Quotas track read, write, and search operations separately for FHIR, and DICOMweb operations for DICOM. Request size limits are enforced per store type. Numeric quota thresholds are not published in documentation and are only visible through the Google Cloud Console Quotas page; projects with high-traffic workloads must proactively request quota increases. FHIR transaction bundles exceeding 4,500 entries are immediately rejected. The API can scale to thousands of requests per second when sufficient quota is allocated.

Google Cloud Healthcare Rate Limits is the machine-readable rate-limit profile for Google Cloud Healthcare on the APIs.io network, conforming to the API Commons Rate Limits specification.

It captures 13 rate-limit definitions, measuring operations_per_minute, entries_per_bundle, and bytes.

The profile also includes 6 backoff/retry policies defined and response codes documented for throttled, quotaExceeded, and serverError.

Tagged areas include Healthcare, FHIR, HL7v2, DICOM, and Google Cloud.

13 Limits Throttle: 429 Quota: 429
HealthcareFHIRHL7v2DICOMGoogle CloudRate Limiting

Limits

FHIR read operations quota project/region
operations_per_minute · minute
see Cloud Console quotas page
Tracks individual FHIR resource reads per project per region. Viewable in GCP Console.
FHIR write operations quota project/region
operations_per_minute · minute
see Cloud Console quotas page
Tracks FHIR create, update, and delete operations per project per region.
FHIR search operations quota project/region
operations_per_minute · minute
see Cloud Console quotas page
Tracks FHIR search requests per project per region.
FHIR transaction bundle max entries request
entries_per_bundle
4500
Transaction bundles with more than 4,500 entries are immediately rejected to prevent timeouts.
FHIR batch bundle timeout risk request
entries_per_bundle
see documentation
Batch bundles have no hard entry limit, but large bundles risk server-side timeouts.
DICOMweb operations quota project/region
operations_per_minute · minute
see Cloud Console quotas page
Tracks STOW-RS, WADO-RS, and QIDO-RS DICOMweb operations per project per region.
HL7v2 message size limit request
bytes
10485760
Maximum HL7v2 message size is 10 MB (base64-encoded data field).
FHIR request size limit (general) request
bytes
10485760
Maximum request body size for most FHIR methods is 10 MB.
FHIR Binary resource size limit request
bytes
1073741824
FHIR Binary resource operations allow up to 1 GB request size.
FHIR executeBundle size limit request
bytes
52428800
FHIR executeBundle (batch/transaction) requests are limited to 50 MB.
DICOM store transaction size limit request
bytes
-1
DICOM STOW-RS store transactions are unlimited in size; import from Cloud Storage is limited only by GCS 5 TB per object maximum.
DICOM non-store request size limit request
bytes
10485760
Non-STOW-RS DICOM methods (WADO-RS retrieve, QIDO-RS query) are limited to 10 MB per request.
Import operation content size operation
bytes
5497558138880
Bulk import operations via Cloud Storage allow content up to the GCS per-object maximum of 5 TB.

Policies

Quota allocation
Default quotas are assigned per project and are not automatically scaled. Projects expecting high-volume healthcare data pipelines must proactively request quota increases via the Google Cloud Console Quotas page before onboarding production traffic.
Quota increase requests
Quota raise requests are submitted through the GCP Console. Allow 1-3 business days for review and approval. For large-scale workloads, contact Google Cloud healthcare sales to coordinate quota planning.
Exponential backoff
Clients receiving 429 (quota exceeded) or 503 (service unavailable) responses should implement exponential backoff with jitter before retrying. Do not retry immediately on quota errors.
FHIR bundle sizing
Keep FHIR transaction bundles under 4,500 entries to avoid immediate rejection. For large data loads, prefer batch bundles or bulk import from Cloud Storage over large transaction bundles.
DICOM ingestion throughput
For high-volume DICOM ingestion, use bulk import from Cloud Storage rather than STOW-RS per-instance upload. This avoids per-request overhead and leverages Cloud Storage's high-throughput transfer capabilities.
Regional quota isolation
Quotas are enforced per project per region. Traffic spread across multiple regions does not share quota pools; each region has its own independent quota allocation.

Sources