Add getQueueUrls and SQS queue name#11151
Add getQueueUrls and SQS queue name#11151kaiyan-sheng merged 17 commits intoelastic:masterfrom kaiyan-sheng:fix_s3_sqs
Conversation
kaiyan-sheng
changed the title
kaiyan-sheng
added
Metricbeat
|
Will this also need a backport? |
ruflin
left a comment
ruflin
left a comment
There was a problem hiding this comment.
- Do you know why sometimes they don't have a timestamp?
How are the scaled_float changes related to the PR description? Perhaps easier to review if multiple PR's?
If we have 7.x branch opened, then this will need a backport to 7.x. Because the changes are not related to ec2 metricset, so no backport to 7.0. |
kaiyan-sheng
changed the title
There was a problem hiding this comment.
In general I would advice to not have for loops inside for loops especially if they contain quite a bit of code and instead use functions/methods. This gives also the opportunity to give some nice names to the functions and potentially have docs there. This will also help with testability.
There was a problem hiding this comment.
Good point, I don't know in this case if helped or not :-) I moved the for loop into createSQSEvents. Thanks!
kaiyan-sheng
added
review
and removed
[zube]: In Progress
in progress
|
This will need a rebase. |
| description: > | ||
| TThe number of messages in the queue that are delayed and not available for reading immediately. | ||
| - name: messages.not_visible | ||
| - name: message.not_visible |
There was a problem hiding this comment.
I assume this means we don't check today if the fields are documented? I hope we get this coverage when we have the new "data tests".
ruflin
left a comment
ruflin
left a comment
There was a problem hiding this comment.
LGTM
I think this does not need a changelog as we didn't release it yet?
|
Yeah no changelog :-) |
3 participants
While trying to create Kibana Dashboards for SQS, I realized there are two bugs in the code:
sqs metricset is reporting metrics without queue name, which is needed for dashboards.
When there are more than 1 queue in the same region, events get overwritten and only the last queue get reported. This bug is fixed in this PR by querying ListQueues first to get a list of queues and then loop through each queue name to construct and report events.