Support consumer groups for which member information is unavailable.

[lilyevsky]

This is a fix for the issue #126 .
It allows to report lag for the consumer groups that don't have members information.
Such situation is actually happening for the consumers running under Flink.

Also minor fix in setting properties, KafkaClient.scala lines 61 and 73:
the documentation for java.util.Properties suggests not to use putAll method because it is not type-safe (in fact, in my environment those two lines did not compile). They suggest using setProperty method.

I tested this fix in my local environment.

[lightbend-cla-validator]

At least one pull request committer is not linked to a user. See https://help.github.com/en/articles/why-are-my-commits-linked-to-the-wrong-user#commits-are-not-linked-to-any-user

[yazgoo]

Hi @lilyevsky . I'm having the same issues with flink and simple consumer groups
#124 is a try at fixing that, but maybe your implementation is better, wdyt @seglo ?

[seglo]

Kafka 2.4.1 will be released shortly, which fixes a different issue in Kafka Lag Exporter. I plan to cut a release then, so let's pick one of these solutions soon.

[seglo]

This PR looks a little more succinct than #124, but I like that #124 gives this type of consumer a name, "simple consumers". Let's use that here.

It would be nice to have a test for this. Can either of you (@yazgoo or @lilyevsky) volunteer to write one?

[lightbend-cla-validator]

At least one pull request committer is not linked to a user. See https://help.github.com/en/articles/why-are-my-commits-linked-to-the-wrong-user#commits-are-not-linked-to-any-user

[aklimko]

Any update on this? 🤔

[seglo]

@lilyevsky Do you mind signing the CLA so I can merge this PR? It might be necessary to recreate the PR to retrigger the check.

[lilyevsky]

@seglo I will work on the CLA. Need to do the corporate one, hopefully can be done in a couple of days.

[lilyevsky]

I am waiting for CLA approval by our Legal department, that may take some time. Hopefully within a week I will have it.

[seglo]

Thanks for the update. I may cut a release in the meantime, but we can push another one once your PR is merged.

[seglo]

I released 0.6.1. We can do another release when you get approval. Or if you like I could implement this myself.

[lilyevsky]

@seglo I got the CLA, will upload it shortly

[lilyevsky]

@seglo I just sent an email with the attached CLA, CCd to you

[lilyevsky]

@seglo I confirmed the CLA with my GitHub account and I got the email response from [email protected]

[lightbend-cla-validator]

At least one pull request committer is not linked to a user. See https://help.github.com/en/articles/why-are-my-commits-linked-to-the-wrong-user#commits-are-not-linked-to-any-user

[JustinPihony]

Hi @lilyevsky this is now failing because your github account does not have your email attached to it. You can see that by looking at the commits and how your username is not clickable to your account. Would you be able to add your email so that github can match you accordingly, and we can then re-run the CLA validator?

[lilyevsky]

Hi @JustinPihony I just edited my profile, added both primary email [email protected] and my work email [email protected]. I see my username is now clickable.
Please try running CLA validator again.

[JustinPihony]

Great! You’re all set. I’ll leave it to @seglo or another maintainer to merge :) thanks!

[seglo]

Thanks @JustinPihony

[jweibel22]

Hi @seglo, I'm also waiting for this feature, what's the timeline for the next release?

[s2490]

Hi @seglo is this fix is included in 0.6.2?
Asking because we have upgraded to latest version and issue is still there 🤔

[lilyevsky]

@s2490 I tested 0.6.2 this morning and it works for me. Maybe your issue is different?

[s2490]

@lilyevsky that could be the case, from CLI perspective example lags looks as follows:

TOPIC           PARTITION  CURRENT-OFFSET  LOG-END-OFFSET  LAG             CONSUMER-ID                                   HOST            CLIENT-ID
topic001        0          74893           172550          97657           client002-4c82c1e2-c868-491b-a22f-751914058fa4 /172.20.52.117  client002
topic001        1          117689          172511          54822           client002-4c82c1e2-c868-491b-a22f-751914058fa4 /172.20.52.117  client002
topic001        2          21398           172639          151241          client002-4c82c1e2-c868-491b-a22f-751914058fa4 /172.20.52.117  client002
topic002        0          49740           50964           1224            -                                             -               -
topic002        2          49898           51166           1268            -                                             -               -
topic002        1          49663           50870           1207            -                                             -               -

Please note that some client details are available, others not. For such cases exporter isn't reporting back metrics for topic002 at all, on our environments such situations are pretty common because of custom partition assignment usage.

Surprisingly when all client details are unavailable exporter is showing correct values, so my example could be some kind of corner case, wdyt?

[s2490]

@lilyevsky have you tested my case maybe or I should create dedicated issue for it?