Cache invalidation strategies in Rest API

Introduction

Cache invalidation strategies help keep stored data fresh and correct. They decide when to update or remove cached data so users get the latest information.

When your app stores data temporarily to speed up responses but the data changes often.

When you want to avoid sending outdated information to users in a web or mobile app.

When your server caches API responses but the underlying data updates regularly.

When you want to reduce load on your database by caching but still keep data accurate.

When you build a REST API that serves data which can be updated by users or other systems.

Syntax

Rest API

There is no single code syntax for cache invalidation because it depends on the strategy used. Common strategies include:

1. Time-based expiration:
   Set a time limit (TTL) after which cached data is removed.

2. Manual invalidation:
   Explicitly clear or update cache when data changes.

3. Event-based invalidation:
   Invalidate cache when certain events happen (like data update).

4. Versioning:
   Use version numbers or keys to identify fresh data.

Cache invalidation is often implemented with headers, cache keys, or API logic.

Choosing the right strategy depends on how often your data changes and how fresh it needs to be.

Examples

This HTTP header tells caches to keep data for 60 seconds before invalidating it (time-based expiration).

Rest API

Cache-Control: max-age=60

This API call manually removes cached data for user 123 (manual invalidation).

Rest API

DELETE /cache/user/123

When product 456 changes, the cache for it is cleared (event-based invalidation).

Rest API

On data update event:
  Invalidate cache key 'product_456'

Changing the cache key version forces clients to get fresh data (versioning).

Rest API

Use cache key 'v2_products_list' instead of 'products_list'

Sample Program

This simple Flask app caches data for 5 seconds. If you request /data within 5 seconds, it returns cached data. After 5 seconds, it fetches fresh data. You can clear the cache manually by calling /invalidate.

Rest API

from flask import Flask, jsonify, request
import time

app = Flask(__name__)

cache = {}
CACHE_TTL = 5  # seconds

@app.route('/data')
def get_data():
    current_time = time.time()
    cached = cache.get('data')
    if cached and current_time - cached['time'] < CACHE_TTL:
        return jsonify({'source': 'cache', 'data': cached['value']})
    # Simulate fresh data fetch
    fresh_data = {'message': 'Hello, world!', 'timestamp': current_time}
    cache['data'] = {'value': fresh_data, 'time': current_time}
    return jsonify({'source': 'fresh', 'data': fresh_data})

@app.route('/invalidate')
def invalidate_cache():
    cache.pop('data', None)
    return jsonify({'status': 'cache cleared'})

if __name__ == '__main__':
    app.run(debug=True)

Important Notes

Cache invalidation is tricky but important to avoid showing old data.

Time-based expiration is simple but may serve stale data briefly.

Manual or event-based invalidation gives more control but needs extra code.

Summary

Cache invalidation keeps cached data fresh and accurate.

Common strategies: time-based, manual, event-based, and versioning.

Choose a strategy based on how often your data changes and how fresh it must be.

Practice

1. Which cache invalidation strategy automatically removes cached data after a set period of time?

easy

A. Manual invalidation

B. Event-based invalidation

C. Time-based invalidation

D. Versioning invalidation

Solution

1. Step 1: Understand time-based invalidation

   Time-based invalidation removes cached data after a fixed time, ensuring data freshness automatically.

2. Step 2: Compare with other strategies

   Manual invalidation requires user action, event-based depends on data changes, and versioning uses data versions.

3. Final Answer:

   Time-based invalidation -> Option C

4. Quick Check:

   Automatic removal after time = Time-based invalidation [OK]

Hint: Time-based means cache expires after set time [OK]

Common Mistakes:

• Confusing manual with automatic removal
• Thinking event-based triggers on time
• Mixing versioning with time expiry

2. Which of the following is the correct way to manually invalidate a cache entry in a REST API?

easy

A. Send a PUT request to update cached data

B. Send a GET request to the cache endpoint

C. Send a POST request to retrieve cached data

D. Send a DELETE request to the cache endpoint

Solution

1. Step 1: Identify HTTP method for deletion

   DELETE requests are used to remove resources, including cache entries.

2. Step 2: Understand other methods

   GET retrieves data, POST creates, and PUT updates; none remove cache.

3. Final Answer:

   Send a DELETE request to the cache endpoint -> Option D

4. Quick Check:

   DELETE removes cache entry manually [OK]

Hint: DELETE method removes resources like cache entries [OK]

Common Mistakes:

• Using GET to remove cache
• Confusing POST with deletion
• Thinking PUT deletes cache

3. Consider this pseudo-code for event-based cache invalidation:

if (dataChangedEvent) {
  invalidateCache(key);
}
return getCachedData(key);

What will happen if dataChangedEvent is true?

medium

A. Cache is invalidated and fresh data is fetched next time

B. Cache remains unchanged and stale data is returned

C. Cache is updated without invalidation

D. An error occurs because cache cannot be invalidated

Solution

1. Step 1: Analyze the condition

   If dataChangedEvent is true, invalidateCache(key) runs, removing cached data.

2. Step 2: Understand cache behavior after invalidation

   After invalidation, next data fetch will get fresh data, not stale cache.

3. Final Answer:

   Cache is invalidated and fresh data is fetched next time -> Option A

4. Quick Check:

   Event triggers invalidation = fresh data next fetch [OK]

Hint: Event true means cache cleared for fresh data [OK]

Common Mistakes:

• Assuming cache stays after invalidation
• Thinking cache updates automatically
• Believing invalidation causes errors

4. You have this code snippet for versioning cache invalidation:

cache = { 'v1': data1 }
function getData(version) {
  return cache[version];
}

// Later
cache['v2'] = newData;

What is the main bug if the client still requests v1 after update?

medium

A. Cache throws an error because version key is missing

B. Client gets stale data because old version is still cached

C. Cache automatically updates old version to new data

D. Client cannot access any cached data

Solution

1. Step 1: Understand versioning cache behavior

   Old version 'v1' remains in cache, so requests for it get stale data.

2. Step 2: Identify bug impact

   Client requesting 'v1' does not see 'v2' data, causing stale responses.

3. Final Answer:

   Client gets stale data because old version is still cached -> Option B

4. Quick Check:

   Old version key returns stale data [OK]

Hint: Old version keys keep stale data unless removed [OK]

Common Mistakes:

• Assuming cache auto-updates old versions
• Expecting errors on missing keys
• Thinking client loses all cache access

5. You want to design a cache invalidation strategy for a REST API where data changes frequently but must stay fresh within 5 minutes. Which combined strategy is best?

hard

A. Use time-based invalidation with 5-minute expiry plus event-based invalidation on data changes

B. Use manual invalidation only, requiring user to clear cache

C. Use versioning only without time limits

D. Use time-based invalidation with 1-hour expiry only

Solution

1. Step 1: Analyze freshness requirement

   Data must be fresh within 5 minutes, so time-based expiry should be 5 minutes or less.

2. Step 2: Combine with event-based invalidation

   Event-based invalidation clears cache immediately on data changes, improving freshness.

3. Step 3: Evaluate other options

   Manual invalidation is slow, versioning alone may keep stale data, and 1-hour expiry is too long.

4. Final Answer:

   Use time-based invalidation with 5-minute expiry plus event-based invalidation on data changes -> Option A

5. Quick Check:

   Combine time and event for fresh cache [OK]

Hint: Combine time and event for best freshness [OK]

Common Mistakes:

• Relying only on manual invalidation
• Using too long expiry times
• Ignoring event triggers for changes