feat(kafka-observ-lib): Expand kafka-observ-lib alerts/panels descriptions and add new alerts

[v-zhuravlev]

Here is what has been updated:
Fix typo in totalTime

Fix typo in preferredReplicaImbalance

Increase size of status rows

Update signals/panels descriptions

Fix typo in KafkaOfflinePartitionCount alert

Add new alerts

Expanded alerts descriptions with runbooks

[Dasomeone]

Apologies on the delay of this review, I will try to get to it tomorrow :)

[Dasomeone]

Nice, thank you @v-zhuravlev, apologies for the delay on this one I kept getting distracted with other things, but all good from my end here :D

[aalhour]

Thank you for improving this, I have two notes about two alerts:

• KafkaOfflinePartitionCount
• KafkaUnderMinISRPartitionCount

Happy to chat about them in the comments.

[aalhour]

If I am not mistaken, the offline partitions in Kafka usually stem from instances (brokers), do you also want to group by the instance here? The group labels don't include it.

I see that the alert KafkaUnderReplicatedPartitionCount nicely includes the instance labels on line 168:

std.join(',', this.config.groupLabels + this.config.instanceLabels),

[aalhour]

I don't have data to prove this out of the box but it seems to me that 2m might short to account for fluctuations that normal operations might cause. Do you agree? I think 5m that is similar to other alerts is appropriate but still short enough for a critical alert to fire.