blacklocus/queue-slayer

(experimental) Your queues -- slay them all.     ☹ ☹⚔qqq Qqqqq⚔☹   ☹⚔q q q' ⇽⤛   ♘☹

queue-slayer's Issues

introduce notion of sampling for log volume control

only track every {predicate} tasks, maybe a decorator around a QSLogService.

predicates something like

• every n tasks
• at most once per minimum interval

Any plans to support "raw", or free form, message processing?

Howdy 🎉

I like what you've done here 👍 and would personally like to use this library to process "raw" messages in various JSON formats. Any thoughts on making the code support arbitrary message formats? I'm primarily referring to not requiring messages to conform to the format imposed by QSTaskModel.

apply current batchId to derivative tasks

Make the `OriginatingMessage` inner class public

Within the AmazonSQSPrioritizedMessageProvider is an inner OriginatingMessage class that is used to wrap a vanilla Message object. This allows for an originatingQueueUrl to be part of the message. Very handy. However the inner class is package-scope so we're not able to take advantage of this value to get the originating URL.

Make it public, please.

augment task object with triggering task

Add field to task model indicating which task triggered this one. Root tasks would have no triggering task. We can then navigate task chains.

add weighted message provider

in cases where we want to guarantee some minimum trickle rate for a number of queues

elasticsearch log service impl fails to log retrying tasks

There is a hard check in logging a "started task" event to not overwrite an existing record. For tasks that are actually retried, this results in a colliding document.

If we are not using ES-generated IDs, consider using those. Or append retry to the taskId to make it unique.

replace all jackson usages with fasterxml version