Retry settings

sparqlite includes automatic retry with exponential backoff for transient errors. This helps your application handle temporary failures gracefully.

Default behavior

By default, sparqlite will:

• Retry up to 5 times on transient errors
• Use exponential backoff with a factor of 0.5

from sparqlite import SPARQLClient

# Default settings: max_retries=5, backoff_factor=0.5
with SPARQLClient("https://dbpedia.org/sparql") as client:
    result = client.query(query)

Configuring retries

max_retries

Set the maximum number of retry attempts:

# Retry up to 3 times
with SPARQLClient("https://dbpedia.org/sparql", max_retries=3) as client:
    result = client.query(query)

# Disable retries entirely
with SPARQLClient("https://dbpedia.org/sparql", max_retries=0) as client:
    result = client.query(query)

backoff_factor

Control the delay between retries. The wait time is calculated as:

wait_time = backoff_factor * (2 ^ attempt)

Example with backoff_factor=0.5:

• Attempt 1: wait 1 second (0.5 * 2^1)
• Attempt 2: wait 2 seconds (0.5 * 2^2)
• Attempt 3: wait 4 seconds (0.5 * 2^3)
• Attempt 4: wait 8 seconds (0.5 * 2^4)
• Attempt 5: wait 16 seconds (0.5 * 2^5)

# Faster retries
with SPARQLClient("https://dbpedia.org/sparql", backoff_factor=0.25) as client:
    result = client.query(query)

# Slower retries
with SPARQLClient("https://dbpedia.org/sparql", backoff_factor=1.0) as client:
    result = client.query(query)

What triggers retry

sparqlite automatically retries on these transient errors:

Error type	Description
HTTP 5xx	Server errors (500, 502, 503, etc.)
Connection failure	Cannot connect to the endpoint
Timeout	Request timed out

What does NOT trigger retry

These errors fail immediately without retry:

Error type	Description
HTTP 400	Query syntax error (QueryError)
HTTP 4xx	Client errors (401, 403, 404, etc.)
Parse error	Invalid response format