Producing and consuming messages tw review #197
Conversation
Consuming-Messages.md
Outdated
| Karafka framework has a long-running server process responsible for fetching and consuming messages. Consumers should inherit from the **ApplicationConsumer**. You need to define a ```#consume``` method that will execute your business logic code against a batch of messages. Karafka fetches and consumes messages in batches by default. | ||
|
|
||
| Karafka fetches and consumes messages in batches by default. | ||
| ### Consuming Messages in Batches |
There was a problem hiding this comment.
my mistake here, it will be changed to H2
Consuming-Messages.md
Outdated
| ``` | ||
| **Result:** The `#consume_one` method will be called for each message in the batch, allowing you to process messages individually while maintaining the benefits of Karafka's batch fetching. | ||
|
|
||
| ### Accessing Topic Details |
There was a problem hiding this comment.
Currently, in Wiki it is H3 displayed under the Consuming Messages Topic. Is it a subsection? Or can I move it/treat it as H2 in TOC? I understood, it is a separate procedure from Consuming Messages (in batches or individually) per se.
| @@ -1,22 +1,33 @@ | |||
| Consumers should inherit from the **ApplicationConsumer**. You need to define a ```#consume``` method that will execute your business logic code against a batch of messages. | |||
| Karafka framework has a long-running server process responsible for fetching and consuming messages. Consumers should inherit from the **ApplicationConsumer**. You need to define a ```#consume``` method that will execute your business logic code against a batch of messages. Karafka fetches and consumes messages in batches by default. | |||
There was a problem hiding this comment.
What do you mean by "define a consume method"? Where is it done/shown? Is it a step?
I wrote: Select one of two processing approaches based on your use case:
Is it when a consume method is defined?
| With [Long-Running Jobs](Pro-Long-Running-Jobs), the `#revoked?` method still updates automatically, even when messages remain unmarked for extended periods. | ||
|
|
||
|
|
||
| ## Consumer Persistence |
There was a problem hiding this comment.
Would you like to rename it to "Buffering Messages with Consumer Persistence" - this signals both the technique (buffering) and the enabling mechanism (persistence).
What's your preference? Do you want developers to understand the broader concept, or focus on the buffering pattern?
| This configuration ensures that as soon as the end of a partition is reached, any accumulated messages are immediately processed, enhancing the system responsiveness and efficiency. | ||
|
|
||
| ## Inline API Based Consumption | ||
| ## Inline API-Based Consumption |
There was a problem hiding this comment.
Better alternatives that align with Karafka's style:
"Iterator API" (simple, feature-focused). It matches the pattern of other feature sections.
"Consuming with the Iterator API" (procedural gerund).
@mensfeld let me know which title is better?
1 participant
No description provided.