Concurrency & Performance
Optimize Text2EverythingClient
for high-throughput and long-running requests.
Recommended settings
client = Text2EverythingClient(
base_url="https://...",
api_key="...",
read_timeout=300,
max_connections=100,
max_keepalive_connections=20,
keepalive_expiry=300.0,
)
Tips
- Prefer session reuse; avoid recreating the client per request.
- Use
chat_to_sql
for generation-only;chat_to_answer
when you need execution. - For heavy result sets, add limits in your SQL to reduce transfer time.
- Monitor 429s and backoff using
retry_after
.
Long-running tasks
- Increase
read_timeout
appropriately. - Ensure downstream DB timeouts (e.g., warehouse) are compatible.
Feedback
- Submit and view feedback for this page
- Send feedback about Text 2 Everything Python Documentation to cloud-feedback@h2o.ai