Time-based event queues in Redis - Time & Space Complexity

Time Complexity: Time-based event queues

O(log n + m)

Understanding Time Complexity

When using Redis for time-based event queues, it is important to understand how the time to process events grows as more events are added.

We want to know how the number of operations changes when the queue gets bigger.

Scenario Under Consideration

Analyze the time complexity of the following Redis commands used to manage a time-based event queue.

ZADD event_queue 1650000000 "event1"
ZADD event_queue 1650000050 "event2"
ZADD event_queue 1650000100 "event3"

ZRANGEBYSCORE event_queue -inf 1650000050
ZREM event_queue "event1"
    

This code adds events with timestamps to a sorted set, retrieves events up to a certain time, and removes processed events.

Identify Repeating Operations

Look at what repeats when managing the queue.

• Primary operation: Retrieving events with ZRANGEBYSCORE which scans events by score (time).
• How many times: This happens every time we want to process events up to a certain time.

How Execution Grows With Input

As the number of events (n) grows, the time to add an event stays quite stable, but retrieving events depends on how many events match the time range.

Input Size (n)	Approx. Operations for ZRANGEBYSCORE
10	About 10 or fewer operations
100	About 100 or fewer operations
1000	About 1000 or fewer operations

Pattern observation: The retrieval cost grows roughly in proportion to the number of events returned.

Final Time Complexity

Time Complexity: O(log n + m)

This means it takes a small time to find where to start, plus time proportional to the number of events returned.

Common Mistake

[X] Wrong: "Retrieving events from the queue always takes the same time no matter how many events there are."

[OK] Correct: The time depends on how many events match the time range, so more events means more work.

Interview Connect

Understanding how Redis handles time-based event queues helps you explain efficient event processing in real systems.

Self-Check

"What if we changed from ZRANGEBYSCORE to using a Redis stream for events? How would the time complexity change?"

Practice

1. What is the main purpose of using a Redis sorted set for a time-based event queue?

easy

A. To store events in a hash for quick lookup

B. To store events with timestamps as scores for ordered retrieval

C. To store only unique event names without timestamps

D. To store events without any order

Solution

1. Step 1: Understand Redis sorted sets

   Redis sorted sets store members with a score that determines their order.

2. Step 2: Apply to time-based event queues

   Using timestamps as scores allows events to be retrieved in chronological order.

3. Final Answer:

   To store events with timestamps as scores for ordered retrieval -> Option B

4. Quick Check:

   Sorted sets + timestamps = ordered event queue [OK]

Hint: Remember: sorted sets order by score (timestamp) [OK]

Common Mistakes:

• Thinking sorted sets don't keep order
• Confusing hashes with sorted sets
• Ignoring the timestamp as score

2. Which Redis command correctly adds an event with a timestamp score to a sorted set named events?

easy

A. SADD events 1625097600 "event1"

B. HSET events 1625097600 "event1"

C. ZADD events 1625097600 "event1"

D. LPUSH events 1625097600 "event1"

Solution

1. Step 1: Identify the command for sorted sets

   ZADD adds members with scores to sorted sets; others are for different data types.

2. Step 2: Check syntax correctness

   ZADD events 1625097600 "event1" adds "event1" with score 1625097600 to sorted set events.

3. Final Answer:

   ZADD events 1625097600 "event1" -> Option C

4. Quick Check:

   ZADD = add to sorted set [OK]

Hint: Use ZADD for sorted sets, not SADD or HSET [OK]

Common Mistakes:

• Using SADD which is for sets without scores
• Using HSET which is for hashes
• Using LPUSH which is for lists

3. Given the sorted set events with members and scores:
{"event1": 1000, "event2": 1500, "event3": 2000}
What will the command ZREVRANGEBYSCORE events 1500 -inf return?

medium

A. ["event1"]

B. ["event3", "event2", "event1"]

C. ["event1", "event2"]

D. ["event2", "event1"]

Solution

1. Step 1: Understand ZREVRANGEBYSCORE command

   This command returns members with scores between max and min in descending order.

2. Step 2: Apply to given scores

   Scores ≤ 1500 are event1 (1000) and event2 (1500). Descending order: event2, event1.

3. Final Answer:

   ["event2", "event1"] -> Option D

4. Quick Check:

   Descending scores ≤ 1500 = event2, event1 [OK]

Hint: ZREVRANGEBYSCORE returns descending order by score [OK]

Common Mistakes:

• Confusing ascending with descending order
• Including events with scores above max
• Misreading score boundaries

4. You run this Redis command to fetch due events:
ZREVRANGEBYSCORE events 1625097600 -inf
But you notice the same events are returned repeatedly. What is the likely mistake?

medium

A. Not removing processed events from the sorted set

B. Using ZREVRANGEBYSCORE instead of ZRANGEBYSCORE

C. Using the wrong key name for the sorted set

D. Using a timestamp that is too high

Solution

1. Step 1: Understand event queue behavior

   After processing events, they must be removed to avoid duplicates.

2. Step 2: Identify cause of repeated events

   If events are not removed, the same events appear on every fetch.

3. Final Answer:

   Not removing processed events from the sorted set -> Option A

4. Quick Check:

   Remove events after processing to avoid repeats [OK]

Hint: Always delete processed events from sorted set [OK]

Common Mistakes:

• Thinking order command causes repeats
• Ignoring event removal step
• Assuming wrong key name causes repeats

5. You want to implement a Redis time-based event queue that fetches all events due up to now and removes them atomically to avoid duplicates. Which sequence of commands is best?

hard

A. Use ZRANGEBYSCORE to get events ≤ now, then ZREM to remove them

B. Use ZREVRANGEBYSCORE to get events ≥ now, then ZREM to remove them

C. Use ZRANGEBYSCORE with a future timestamp, then LPUSH to add events

D. Use SPOP to pop events from a set

Solution

1. Step 1: Fetch due events with ZRANGEBYSCORE

   ZRANGEBYSCORE with max = current timestamp fetches all events due up to now.

2. Step 2: Remove fetched events with ZREM

   Removing them prevents duplicates; doing both steps atomically avoids race conditions.

3. Final Answer:

   Use ZRANGEBYSCORE to get events ≤ now, then ZREM to remove them -> Option A

4. Quick Check:

   Fetch ≤ now + remove = atomic event processing [OK]

Hint: Fetch with ZRANGEBYSCORE ≤ now, then remove with ZREM [OK]

Common Mistakes:

• Using ZREVRANGEBYSCORE with wrong score range
• Using sets or lists instead of sorted sets
• Not removing events after fetching