Reduce The Cost Of Your Lambda functions With Event Filters(Part 2 of 2)

This cheat sheet covers the advanced syntax for AWS Lambda event filtering, allowing you to build more complex and precise rules to control which events trigger your function. This is a continuation of the basic event filtering concepts.

Filter Rule Syntax

Lambda event filter rules are defined in a JSON object. The structure of the JSON must match the structure of the event payload you want to filter. The filter engine supports a rich set of comparison operators.

Basic Filtering (Recap)

Exact Match:
- Payload: { "status": "completed" }
- Filter: { "status": ["completed"] }
OR Logic (Any value in the array):
- Payload: { "priority": "high" }
- Filter: { "priority": ["high", "critical"] } (Triggers if priority is "high" OR "critical")

Advanced Filtering Rules

Numeric Matching

You can filter on numeric fields using comparison operators.

Rule:

{ "price": [{ "numeric": [">", 0, "<=", 150.50] }] }

Matches:
- { "price": 100 }
- { "price": 150.50 }
Does NOT Match:
- { "price": 0 }
- { "price": 200 }

`exists` - Checking for Presence

The exists operator checks whether a specific field is present in the payload.

Rule:

{ "customer": { "email": [{ "exists": true }] } }

Matches: { "customer": { "email": "test@example.com" } }
Does NOT Match: { "customer": { "phone": "555-1234" } } (The email key is missing)

You can also check for non-existence:

Rule:

{ "customer": { "email": [{ "exists": false }] } }

`prefix` - Matching the Beginning of a String

The prefix operator checks if a string value starts with a specific substring.

Rule:

{ "eventType": [{ "prefix": "order_" }] }

Matches:
- { "eventType": "order_created" }
- { "eventType": "order_shipped" }
Does NOT Match: { "eventType": "payment_processed" }

`anything-but` - Exclusion Logic

The anything-but operator matches any value except for the ones specified. This is useful for ignoring certain event types.

Rule:

{ "state": [{ "anything-but": ["archived", "deleted"] }] }

Matches:
- { "state": "pending" }
- { "state": "active" }
Does NOT Match:
- { "state": "archived" }
- { "state": "deleted" }

You can also use anything-but with a prefix:

Rule: { "source": [{ "anything-but": { "prefix": "internal-" } }] } (Matches any source that does not start with "internal-")

Real-World Example: DynamoDB Streams

Imagine you have a DynamoDB stream capturing item-level changes. You want to trigger a Lambda function only when a new customer is added with a creditScore over 700.

Sample DynamoDB Stream Record (simplified payload):

{
  "eventName": "INSERT",
  "dynamodb": {
    "NewImage": {
      "customerId": { "S": "cust-9876" },
      "status": { "S": "active" },
      "creditScore": { "N": "750" }
    }
  }
}

Advanced Filter Criteria:

This filter combines multiple rules to achieve the desired outcome.

{
  "eventName": ["INSERT"],
  "dynamodb": {
    "NewImage": {
      "creditScore": [{ "numeric": [">", 700] }],
      "status": [{ "anything-but": ["on_hold"] }]
    }
  }
}

How this filter works:

"eventName": ["INSERT"]: Ensures the function only triggers for newly created items, not for updates or deletions.
"creditScore": [{ "numeric": [">", 700] }]: Applies a numeric filter on the creditScore field within the NewImage part of the payload. Note that DynamoDB Stream values are strings, but the filter engine can evaluate them numerically.
"status": [{ "anything-but": ["on_hold"] }]: Adds an extra condition to ignore new customers who might be in an "on_hold" state.

By using these advanced rules, you can create highly specific triggers, drastically reducing unnecessary Lambda invocations and making your serverless architecture more efficient and cost-effective.

Reduce The Cost Of Your Lambda functions With Event Filters(Part 2 of 2)

📚 Recommended AWS Resources