Lambda concurrency and throttling in AWS - Step-by-Step Execution

Process Flow - Lambda concurrency and throttling

Incoming Lambda Requests

↓

Check Current Concurrency

↓

Below Limit

↓

Invoke Lambda

↓

Process

↓

Complete Execution

↓

Decrease Concurrency Count

Lambda checks how many functions run at once. If below limit, it runs. If at limit, it stops new runs and returns an error.

Execution Sample

AWS

1. Set concurrency limit = 3
2. Receive 5 requests quickly
3. Invoke Lambda if concurrency < 3
4. Throttle if concurrency >= 3

Simulates 5 quick Lambda calls with a concurrency limit of 3 to show which run and which get throttled.

Process Table

Step	Incoming Request #	Current Concurrency	Condition (Concurrency < Limit?)	Action	Result
1	1	0	0 < 3 (True)	Invoke Lambda	Request 1 runs, concurrency = 1
2	2	1	1 < 3 (True)	Invoke Lambda	Request 2 runs, concurrency = 2
3	3	2	2 < 3 (True)	Invoke Lambda	Request 3 runs, concurrency = 3
4	4	3	3 < 3 (False)	Throttle Request	Request 4 throttled, concurrency = 3
5	5	3	3 < 3 (False)	Throttle Request	Request 5 throttled, concurrency = 3
6	Request 1 completes	3	-	Decrease concurrency	Concurrency = 2
7	Request 4 retry	2	2 < 3 (True)	Invoke Lambda	Request 4 runs, concurrency = 3

💡 Requests 4 and 5 are throttled because concurrency reached the limit of 3.

Status Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 4	After Step 5	After Step 6	After Step 7
concurrency	0	1	2	3	3	3	2	3
request_status	none	running(1)	running(1,2)	running(1,2,3)	throttled(4)	throttled(5)	running(2,3)	running(2,3,4)

Key Moments - 2 Insights

Why are requests 4 and 5 throttled even though they arrived after requests 1, 2, and 3?

Because concurrency was already at the limit 3 when requests 4 and 5 arrived (see steps 4 and 5 in execution_table), Lambda cannot run more than 3 at once.

What happens to concurrency when a Lambda request finishes?

Concurrency decreases by 1 (step 6), allowing new requests to run if they retry (step 7).

Visual Quiz - 3 Questions

Look at the execution table, what is the concurrency value after step 3?

A 3

B 2

C 1

D 0

💡 Hint

Check the 'Current Concurrency' column at step 3 in the execution_table.

At which step does the first request get throttled?

A Step 2

B Step 4

C Step 3

D Step 5

💡 Hint

Look for the first 'Throttle Request' action in the execution_table.

If the concurrency limit was increased to 5, what would happen at step 5?

A Request 4 would be throttled

B Request 5 would be throttled

C Request 5 would run

D No requests would run

💡 Hint

Compare concurrency and limit at step 5 in the execution_table and imagine the limit is 5 instead of 3.

Concept Snapshot

Lambda concurrency limits how many functions run at once.
If concurrency reaches the limit, new requests are throttled.
When a function finishes, concurrency decreases.
Throttled requests can retry later.
This protects backend resources from overload.

Full Transcript

This visual execution shows how AWS Lambda handles concurrency and throttling. When Lambda receives requests, it checks how many functions are running. If the number is below the concurrency limit, Lambda runs the function and increases concurrency. If the limit is reached, Lambda throttles new requests, returning an error. When a running function finishes, concurrency decreases, allowing throttled requests to retry and run. This process ensures Lambda does not overload resources by running too many functions at once.

Practice

1. What does Lambda concurrency control in AWS Lambda functions?

easy

A. The total number of functions in an AWS account

B. The number of function instances that can run at the same time

C. The size of the function's deployment package

D. The amount of memory allocated to the function

Solution

1. Step 1: Understand Lambda concurrency concept

   Concurrency in Lambda means how many copies of your function can run simultaneously.

2. Step 2: Identify what concurrency controls

   It controls the number of function instances running at the same time to handle requests.

3. Final Answer:

   The number of function instances that can run at the same time -> Option B

4. Quick Check:

   Concurrency = simultaneous function runs [OK]

Hint: Concurrency = how many functions run at once [OK]

Common Mistakes:

• Confusing concurrency with memory size
• Thinking concurrency limits total functions
• Mixing concurrency with deployment package size

2. Which of the following is the correct way to set reserved concurrency for a Lambda function using AWS CLI?

easy

A. aws lambda update-function --name MyFunc --concurrency 5

B. aws lambda set-concurrency --function MyFunc --reserved 5

C. aws lambda put-function-concurrency --function-name MyFunc --reserved-concurrent-executions 5

D. aws lambda configure-concurrency --function MyFunc --limit 5

Solution

1. Step 1: Recall AWS CLI command for reserved concurrency

   The correct command is 'put-function-concurrency' with '--reserved-concurrent-executions' option.

2. Step 2: Match the correct syntax

   aws lambda put-function-concurrency --function-name MyFunc --reserved-concurrent-executions 5 matches the exact AWS CLI syntax to set reserved concurrency.

3. Final Answer:

   aws lambda put-function-concurrency --function-name MyFunc --reserved-concurrent-executions 5 -> Option C

4. Quick Check:

   Reserved concurrency CLI = put-function-concurrency [OK]

Hint: Use 'put-function-concurrency' with reserved count [OK]

Common Mistakes:

• Using wrong command names like 'set-concurrency'
• Incorrect option flags
• Confusing update-function with concurrency settings

3. Given a Lambda function with a reserved concurrency of 3, what happens if 5 requests arrive simultaneously?

medium

A. 3 requests run immediately; 2 are throttled and may be retried

B. All 5 requests run simultaneously without delay

C. Only 1 request runs; others are queued

D. All 5 requests are rejected immediately

Solution

1. Step 1: Understand reserved concurrency effect

   Reserved concurrency limits how many instances run at once; here it is 3.

2. Step 2: Analyze request handling

   When 5 requests come, 3 run immediately; the other 2 exceed limit and get throttled (rejected or retried).

3. Final Answer:

   3 requests run immediately; 2 are throttled and may be retried -> Option A

4. Quick Check:

   Requests > concurrency = throttling [OK]

Hint: Requests over concurrency limit get throttled [OK]

Common Mistakes:

• Assuming all requests run simultaneously
• Thinking requests queue automatically
• Believing throttled requests are dropped without retry

4. You set a reserved concurrency of 2 for a Lambda function but notice throttling errors even with only 2 requests at once. What is the likely cause?

medium

A. The reserved concurrency setting was not applied correctly

B. The account-level concurrency limit is lower than 2

C. The function timeout is set too high

D. The function's memory size is too small

Solution

1. Step 1: Check reserved concurrency application

   If throttling occurs at 2 requests, reserved concurrency might not be set or applied properly.

2. Step 2: Rule out other causes

   Memory size and timeout do not cause throttling; account limit lower than 2 is unlikely if 2 requests run.

3. Final Answer:

   The reserved concurrency setting was not applied correctly -> Option A

4. Quick Check:

   Throttling with low requests = misconfigured concurrency [OK]

Hint: Check if reserved concurrency is actually set [OK]

Common Mistakes:

• Blaming memory size for throttling
• Ignoring concurrency setting errors
• Assuming timeout causes throttling

5. You have two Lambda functions, A and B, sharing an account concurrency limit of 10. Function A has reserved concurrency 7. If both functions receive 7 simultaneous requests each, how many requests for function B will be throttled?

hard

A. 3 requests for function B will be throttled

B. 0 requests for function B will be throttled

C. 7 requests for function B will be throttled

D. 4 requests for function B will be throttled

Solution

1. Step 1: Understand reserved concurrency reservation

   Function A reserves 7 concurrency, so it can run up to 7 instances without throttling.

2. Step 2: Calculate remaining concurrency for function B

   Account limit is 10, so 10 - 7 = 3 concurrency left for function B.

3. Step 3: Analyze requests for function B

   Function B gets 7 requests but only 3 can run concurrently; 7 - 3 = 4 requests are throttled.

4. Final Answer:

   4 requests for function B will be throttled -> Option D

5. Quick Check:

   Reserved concurrency reduces available concurrency [OK]

Hint: Reserved concurrency reduces shared limit for others [OK]

Common Mistakes:

• Ignoring reserved concurrency reservation
• Assuming full account concurrency is available to all
• Miscounting throttled requests