Token bucket algorithm in Rest API - Step-by-Step Execution

Concept Flow - Token bucket algorithm

Start

↓

Tokens added at fixed rate

↓

Request arrives

↓

Check if tokens >= request cost?

No→Reject request

Yes↓

Consume tokens

↓

Allow request

↓

Wait for next request or token refill

↓

Request arrives

Tokens are added steadily. Each request uses tokens if enough are available; otherwise, it is rejected.

Execution Sample

Rest API

bucket_capacity = 5
tokens = 5
refill_rate = 1  # tokens per second

# Request costs 2 tokens
if tokens >= 2:
    tokens -= 2  # consume tokens
    allow_request = True
else:
    allow_request = False

This code checks if there are enough tokens for a request costing 2 tokens, consumes them if possible, and decides to allow or reject the request.

Execution Table

Step	Time (s)	Tokens Before	Request Cost	Tokens After	Request Allowed
1	0	5	2	3	Yes
2	1	4	3	1	Yes
3	2	3	4	3	No
4	3	4	1	3	Yes
5	4	4	5	4	No
6	5	5	2	3	Yes

💡 Simulation ends after 6 requests processed with token refills each second.

Variable Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 4	After Step 5	After Step 6
tokens	5	3	1	3	3	4	3
allow_request	N/A	Yes	Yes	No	Yes	No	Yes

Key Moments - 3 Insights

Why does the token count increase by 1 before each request?

Tokens are refilled at a fixed rate (1 token per second) before processing each request, as shown in the 'Tokens Before' column in the execution_table.

Why is the request rejected at step 3 even though tokens are available?

At step 3, tokens before request are 3 but request cost is 4, which is more than available tokens, so the request is rejected (see execution_table row 3).

Why does the token count never exceed the bucket capacity?

Tokens refill only up to the bucket capacity (5 tokens). The execution_table shows tokens never go above 5, ensuring the bucket limit is respected.

Visual Quiz - 3 Questions

Look at the execution_table, what is the token count after step 2?

A 4

B 1

C 3

D 2

💡 Hint

Check the 'Tokens After' column in row 2 of the execution_table.

At which step is a request rejected due to insufficient tokens?

A Step 3

B Step 1

C Step 4

D Step 6

💡 Hint

Look for 'No' in the 'Request Allowed' column in the execution_table.

If the refill rate was 2 tokens per second, how would tokens before step 3 change?

A It would be 5

B It would be 3

C It would be 4

D It would be 2

💡 Hint

Tokens refill faster but cannot exceed bucket capacity (5 tokens). Check variable_tracker for token limits.

Concept Snapshot

Token bucket algorithm:
- Tokens added steadily at fixed rate
- Bucket has max capacity
- Each request costs tokens
- If enough tokens, consume and allow
- Else, reject request
- Controls request rate smoothly

Full Transcript

The token bucket algorithm controls how many requests can be processed over time. Tokens are added to a bucket at a steady rate, up to a maximum capacity. Each incoming request needs some tokens. If the bucket has enough tokens, the request is allowed and tokens are consumed. If not, the request is rejected. This way, bursts of requests can be handled smoothly without exceeding limits. The execution table shows tokens before and after each request, whether the request is allowed, and how tokens refill over time.