Details
-
New Feature
-
Status: Open
-
Minor
-
Resolution: Unresolved
-
None
-
None
-
None
Description
By grouping and counting messages by time range, this feature would provide operators with an overview of when messages still present in the queue originally arrived.
The primary use-case would be for dead-letter queues and for investigating problems based on the time ranges highlighted in the returned data.
The output is in JSON, start and end indicates the range, and count is the number of messages found within that range.
{
{
start: "start-timestamp",
end: "end-timestamp",
count: "number-of-messages"
},
...
}
The JSON can be used to generate histograms if needed.
Proposal:
- Method "countMessagesByArrival(groupUnit, filter)".
The groupUnit argument is used to group messages, default value being "Days" (other possibility can be Seconds, Minutes, ...).
A message's group is determined based on its timestamp and the selected groupUnit.
The filter allows messages in the queue to be filtered before the distribution is computed.
Each message in the queue is checked against the filter. The group for each message is determined based on the groupUnit and the message's timestamp.
Finally, a message counter is incremented for the corresponding group.
This method could be expensive, as it needs to scan all messages in a queue. Therefore, the method description should clearly highlight this.
Example output with groupUnit="SECONDS"
{
{
start: "1727706968000",
end: "1727706969000",
count: "3"
},
{
start: "1727706976000",
end: "1727706977000",
count: "10"
},
{
start: "1727706980000",
end: "1727706981000",
count: "10"
}
}