Merge sdk dashboards into one

[antmendoza]

What was changed

• Merged both dashboards into one to make it easier to maintain.

• Some panels have two queries when the metric name for Java/Go and Core SDK does not match. Final users can ignore or delete queries that do not report data for their SDK

• Added local activity metrics

• added temporal_worker_task_slots_used

Why?

Checklist

1. Closes

2. How was this tested:

3. Any docs updates needed?

[kevinawoo]

cool stuff! 👍

[tsurdilo]

i think this does make things a bit easier, but think it will be difficult to be used by actual polyglot microservices.
we do have users (and customers) that use multiple temporal sdks, and some of our sdks (java, go) push timer metrics by default in seconds and others (core sdks) in milliseconds. this i think will be hard to create a single datasource for these graphs where they make sense?

i know its a bit harder to maintain, but to me at least given the differences in metrics by our diff sdk currently doing this is just easier for users as they can just pick the one dashboard that works, instead of seeing a lot of "noice"

[kevinawoo]

it will be difficult to be used by actual polyglot microservices.

In practice though, how often are you going to be looking at timer metrics across all your services at once? I would imagine that it's more useful to see metrics for a single app at a time rather than all at a time. ServiceA is likely not going to perform like ServiceB, even if they're on the same SDK.

[antmendoza]

Thank you both,

My understanding is that this graph is meant to serve as a starting point, and users should/will adapt it according to their needs. I am not sure what users do know when they have run polyglot microservices. Can they add custom labels to the scraper configuration?

and some of our sdks (java, go) push timer metrics by default in seconds and others (core sdks) in milliseconds.

this is a good point. Users can modify the runtime configuration to set 'durations_as_seconds=true' or they can modify the dashboards. Maybe we can add some documentation in the readme.md

[kevinawoo]

Can they add custom labels to the scraper configuration?

I all scrapes in Prometheus will a job_name, so you can identify which service is being scraped. Also in Prometheus, you're able to relable metric labels (add/filter/rename/map, etc...)

[antmendoza]

@kevinawoo @tsurdilo I have created a new pr with two dashboards #60