Google BigQuery · Rate Limits
Google Bigquery Rate Limits
BigQuery enforces a layered set of quotas - per-project concurrent query/job counts, per-user/project query rate limits, table operation ceilings, streaming insert and Storage Write API throughput, and daily caps on certain destructive operations. Most quotas can be raised via Cloud Console quota requests for production workloads. The numbers below are the public defaults from the BigQuery quotas reference; always check the live page for the most current value for your region.
15 Limits
Throttle: 429
Quota: 403
Data WarehouseAnalyticsGoogle CloudSQLRate Limiting
Limits
Concurrent interactive queries project
100
On-demand interactive queries; reservation slots have separate concurrency.
Concurrent legacy SQL queries project
100
Daily query usage project
see quotas page
Configurable custom cost control on bytes billed per day.
API requests per user user/project
100
Load jobs per table per day table
1500
Load jobs per project per day project
100000
Copy jobs per table per day table
100
Copy jobs per project per day project
100000
Export jobs per project per day project
100000
Streaming insert rows per second per project (US/EU multi-region) project
see quotas page
Storage Write API throughput per project (US/EU) project
3000000000
3 GB/s default in US/EU multi-region; can be raised.
Storage Read API simultaneous streams project
see quotas page
Maximum row size (DML/streaming) row
104857600
100 MB.
Maximum table column count table
10000
Table operations per table per day (DML) table
1500
Policies
Quota raises
Most quotas are soft and can be raised via the Google Cloud Console quotas page. Plan a 1-3 day lead time for review.
Backoff
Implement exponential backoff with jitter on 429 / 403 quotaExceeded responses. Reservation-aware clients should respect slot capacity.
Reservation isolation
When using capacity (Standard/Enterprise/Enterprise Plus) reservations, query concurrency is governed by allocated slots rather than the on-demand concurrent-query limit.
Streaming vs Storage Write
Prefer the Storage Write API over the legacy streaming insert API for new work; it has higher throughput, lower cost, and supports exactly-once semantics.