Retry Policies

Guide for creating ServiceNow Retry Policies using the Fluent API. Retry policies control how outbound integration connections recover from transient failures by defining when to retry, how many times, and how long to wait between attempts.

When to Use

• Automatically retry failed requests when a step encounters an intermittent issue such as a network failure or request rate limit, to prevent having to manually trigger the step again
• Outbound HTTP/REST calls that may encounter rate limits (429) or temporary server errors (500, 502, 503, 504)
• JDBC connections that may fail transiently on database restarts or network hiccups
• SFTP connections that require retry logic for reliability

Instructions

1. Choose connectionType first: The connection type determines which retry conditions and strategies are available. Most use cases are HTTP ('http_retry_conditions').
2. Pick the right strategy: 'fixed_time_interval' and 'exponential_backoff' use count and interval; 'retry_after' uses only maxElapsedTime and honors the HTTP Retry-After header.
3. Scope with condition: Use an encoded query on restrictTo fields to limit retries to the right error types. This prevents retrying on non-recoverable errors (e.g., 400 Bad Request).
4. Set restrictTo narrowly: Only include the fields your condition actually uses to avoid evaluating irrelevant conditions.

API Reference

See the retrypolicy-api topic for the full property reference.

Key Concepts

Retry Strategies

fixed_time_interval (default)

Waits exactly interval seconds after each failed attempt before retrying. Use when the downstream service needs a predictable recovery window.

RetryPolicy({
    $id: Now.ID['fixed-retry'],
    retryStrategy: 'fixed_time_interval',
    count: 3,
    interval: 10,
})

exponential_backoff

Exponentially increases the time interval for subsequent retry attempts. The multiplier is 2 — each wait is double the previous. Start with a short interval (e.g., 5 seconds) and let backoff handle growth.

RetryPolicy({
    $id: Now.ID['exponential-retry'],
    retryStrategy: 'exponential_backoff',
    count: 5,
    interval: 5,
})

retry_after

Retries based on the date and time value returned in the Retry-After header of the HTTP response. The header value can be either an HTTP-date or a delay in seconds. Requires maxElapsedTime and is only valid with connectionType: 'http_retry_conditions' (REST steps).

RetryPolicy({
    $id: Now.ID['retry-after'],
    connectionType: 'http_retry_conditions',
    retryStrategy: 'retry_after',
    maxElapsedTime: 300,
})

count and interval are not valid with retry_after — the wait time comes from the response header, not a fixed value.

Condition and restrictTo

Use condition (an encoded query) to limit retries to specific failure scenarios. The restrictTo array declares which fields your condition references — only include fields actually used in the condition.

Encoded query operators:

• IN - Match any value in a list: status_codeIN429,500,502
• ^OR - Logical OR: status_code=429^ORstatus_code=503
• BETWEEN - Range: status_codeBETWEEN500@599
• ISNOTEMPTY - Field has a value: endpoint_errorISNOTEMPTY

// Example 1: Retry on specific status codes or any error message
RetryPolicy({
    $id: Now.ID['selective-retry'],
    connectionType: 'http_retry_conditions',
    retryStrategy: 'exponential_backoff',
    count: 5,
    interval: 5,
    condition: 'status_codeIN429,500,502,503,504^ORendpoint_errorISNOTEMPTY',
    restrictTo: ['status_code', 'error'],
})

// Example 2: Retry on 429 or any 5xx error
RetryPolicy({
    $id: Now.ID['rate-limit-or-5xx'],
    connectionType: 'http_retry_conditions',
    retryStrategy: 'retry_after',
    maxElapsedTime: 300,
    condition: 'status_code=429^ORstatus_codeBETWEEN500@599',
    restrictTo: ['status_code'],
})

condition is only valid for HTTP connections — it is not supported for JDBC or SFTP (basic_retry_conditions).

maxElapsedTime

Sets a ceiling on the total retry window in seconds. Required for retry_after; an error for all other strategies. Must not exceed 86400 (24 hours).

// ✅ Valid: retry_after requires maxElapsedTime
RetryPolicy({
    $id: Now.ID['bounded-retry'],
    retryStrategy: 'retry_after',
    maxElapsedTime: 120,
})

// ✗ Invalid: maxElapsedTime not allowed with fixed_time_interval
RetryPolicy({
    $id: Now.ID['bad-retry'],
    retryStrategy: 'fixed_time_interval',
    count: 3,
    interval: 10,
    maxElapsedTime: 60, // build error
})

Connection Types

Value	Use for
'http_retry_conditions'	REST / outbound HTTP integrations
'jdbc_retry_conditions'	JDBC database connections
'basic_retry_conditions'	SFTP connections

condition and retry_after strategy are only supported for 'http_retry_conditions'.

Examples

Rate-limited API (retry_after)

For APIs like Salesforce or external SaaS that return 429 Too Many Requests with a Retry-After header:

import { RetryPolicy } from '@servicenow/sdk/core'

export const salesforceRetry = RetryPolicy({
    $id: Now.ID['salesforce-retry'],
    name: 'Salesforce Rate Limit Retry',
    connectionType: 'http_retry_conditions',
    retryStrategy: 'retry_after',
    maxElapsedTime: 300,
    condition: 'status_codeIN429',
    restrictTo: ['status_code'],
})

Resilient REST integration (exponential backoff)

For outbound REST calls that may encounter transient server errors:

import { RetryPolicy } from '@servicenow/sdk/core'

export const restRetry = RetryPolicy({
    $id: Now.ID['rest-retry'],
    name: 'REST API Retry',
    connectionType: 'http_retry_conditions',
    retryStrategy: 'exponential_backoff',
    count: 5,
    interval: 5,
    condition: 'status_codeIN500,502,503,504',
    restrictTo: ['http_method', 'status_code', 'error', 'response_body', 'response_headers'],
})

JDBC connection retry

For database connections that may fail during short outages:

import { RetryPolicy } from '@servicenow/sdk/core'

export const dbRetry = RetryPolicy({
    $id: Now.ID['db-retry'],
    name: 'Database Connection Retry',
    connectionType: 'jdbc_retry_conditions',
    retryStrategy: 'fixed_time_interval',
    count: 3,
    interval: 10,
})

Related Topics

• retrypolicy-api — Full property reference for RetryPolicy
• alias-guide — Creating Connection & Credential Aliases that use retry policies