SERVER_UNAVAILABLE: Too Many Requests Waiting for Connections Error in Salesforce

While the Salesforce service is continuously dimensioned to minimize the frequency of this error and ensure high availability, sudden load spikes on a given instance may cause this error to be returned to integrations when database-intensive requests are being processed and connections do not become available in time.

Site Reliability Information

• Salesforce Site Reliability closely monitors service availability on a continuous basis and takes any necessary actions to ensure the frequency of these errors is as low as possible.
• Trust is our #1 priority. a posting will be made at Salesforce Trust  if a confirmed performance degradation takes place on a given instance. However, there may be occasions when no posting is made if the database was under intensive utilization for only a very narrow time window.
• If the error is persistent or experienced with high frequency on a given instance and no posting has been found on Trust, submit a case with Salesforce Support. Support will ensure the issue is investigated with maximum diligence.

Recommended Action: Implement Retry Logic with Exponential Backoff

It is strongly recommended to implement retry logic in your integration to handle these transient errors. The retry logic should use the binary exponential backoff algorithm. Exponential backoff means retrying after increasing wait intervals — for example:

• Retry after 1 second; if it fails again
• Retry after 2 seconds; if it fails again
• Retry after 4 seconds; if it fails again
• Retry after 8 seconds; and so on

This approach minimizes the impact of temporary server load spikes on your integration.