Is this project alive?

[Giovanni Martinez]

I noticed that there is not much activity in GitHub. Is this project alive?

[Peter Corless]

That's a good question. At ScyllaDB we're always happy to hear of new KairosDB use cases where Scylla is being used as the underlying data store.

Who's doing cool work these days with KairosDB? Would love to know. 

As an aside, I saw Brian made some code commits in Nov 2020, but haven't seen much conversation from users. [Will admit I am not always monitoring this channel.]

On Mon, Feb 1, 2021, 8:54 AM Giovanni Martinez <gml...@gmail.com> wrote:

I noticed that there is not much activity in GitHub. Is this project alive?

--
You received this message because you are subscribed to the Google Groups "KairosDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email to kairosdb-grou...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/kairosdb-group/a08e6605-1ccb-45d8-bf18-3897f06e09a4n%40googlegroups.com.

[Loic Coulet]

Project is very quiet.

We still use KairosDB. It's still the best time series database I know about :-)

And using this  Grafana plugin it's probably the most affordable to all kind of users : https://github.com/KC-SAS/kairosdb-grafana-enhanced-plugin

Who's doing cool work these days with KairosDB? I'd also would love to know.

[Peter Corless]

That's a nice plugin!

What backing store are you using for KairosDB?

To view this discussion on the web visit https://groups.google.com/d/msgid/kairosdb-group/245d6808-c448-482e-b513-d719ca3778f7n%40googlegroups.com.

[Brian Hawkins]

It's alive, just taking a nap ;)  I have a release waiting in the wings for me to finish some documentation.  Then I have some big plans for making the query engine much better than it is.

Brian

[Loic Coulet]

Hi Brian,

Good to see you back here.

do you have any insights about the query engine changes you are envisionning ?

Cheers,

Loic

[Dirk Grobler]

Hi Brian, all,

We have been using KairosDB for more than 5 years now for IoT use cases and are happy with it. And I agree, there are not really any better solutions out there.  But the none activity is a concern. There is a new major version of Cassandra around the corner, Iot use cases would be much better served with second granularity, just comes to mind would be good to see on the roadmap. 
We can certainly help to test new releases, if that would make a difference. 

Cheers,

Dirk

[Alexandre Dugarry]

Hi Brian,

Well I am glad that this project is still alive. Like Dirk, we have been using KairosDB for almost 5 years now and I think this is still one of the best solution for timeseries. Very robust, easy to setup and maintain, cheap to run.

Let us know how we can contribute.

Best,

Alex 

[Brian Hawkins]

Dirk,  Second granularity is done: https://github.com/kairosdb/kairosdb/issues/246

Just need to document and release this bad boy.

Loic,  Query engine changes that are in my head.  I think of the query as a directed acyclic graph (DAG).  Each component in the graph is a query component that either does a sum or groups by or any other operation that then has a stream of data coming out of.  The current engine is kind of a limited version of this, what I want to build is one that is more generic where components can easily be swapped around.  So when you create a query you would define the DAG, yes the query language will start out being major ugly, but then I'm thinking that abstractions could be built on top of it.  Ideally a sweet UI that lets you drag and drop components to create the DAG.

That is the rough direction I want to go with it.  It will be a challenge but I'm pretty sure I can do it.

Brian

[Dirk Grobler]

Thanks Brian,

Can't wait to get my hands on it. I assume I will have to setup a new cluster and then replay the old data into the new one, right?

-Dirk

[Brian Hawkins]

Dirk,  Well yes and no.  1.3 supports multi cluster.  So you will have to create a new cluster in order to configure it for 1 sec granularity, but you can have your old cluster as a read cluster - all new data goes into the 1 sec granularity cluster but the old is still accessible via the queries.

Brian