Fixed Delay Retry Strategy
What you'll learn: How to implement retries with a constant wait time between attempts and why this simple approach has important limitations under load.
The Core Idea
A fixed delay retry strategy waits the same amount of time between each retry attempt. If a request fails, you wait (say) 2 seconds, then try again. Fail again? Wait another 2 seconds. Repeat until you succeed or hit your maximum retry limit.
This is the simplest retry strategy to implement:
Attempt 1 → fails
Wait 2 seconds
Attempt 2 → fails
Wait 2 seconds
Attempt 3 → succeeds
Why It's Appealing
Fixed delays are predictable and easy to reason about. You know exactly when the next retry will happen, making debugging and monitoring straightforward. For low-traffic systems or transient network hiccups, this works perfectly fine.
The Problem: The Thundering Herd
Here's where fixed delays fall apart under load. Imagine 1,000 clients all trying to call your overloaded database at exactly 10:00:00 AM. All requests fail. With a fixed 2-second delay, all 1,000 clients retry at exactly 10:00:02 AM. The database gets hammered again, fails again, and now all clients retry at 10:00:04 AM.
You've created a retry storm—synchronized retries that prevent the system from recovering. The constant delay means clients remain synchronized, amplifying the load spike instead of distributing it.
When to Use It
Fixed delay works well for:
- Low-traffic scenarios
- Systems with stable, predictable failure modes
- Quick prototypes
But for production systems at scale, you'll need smarter strategies (like exponential backoff or jitter, which add randomization to break synchronization).
Key Takeaway: Fixed delay retries are simple but dangerous at scale because they can synchronize client behavior and create retry storms that prevent system recovery.