openTSDB 1.1 vs openTSDB 2.0

Showing 1-8 of 8 messages
openTSDB 1.1 vs openTSDB 2.0 bernd 2/26/13 2:38 AM
Hi,

actually i take a closer look at storage solutions for timeseries which came from scada systems.

As i saw both versions (openTSDB 1.1 vs openTSDB 2.0) are actively under development and both versions have cool features for my application.

For example:
openTSDB 1.1 will be able to store annotations respectivly string values for timeseries,
or the really nice feature in openTSDB 2.0 to provide a tree api to preorganise timeseries as a tree.

Especially this both features are very valuable for my  application.

So my question is:
Will openTSDB 2.0 include all upcoming features from openTSDB 1.1?
Is there a roadmap / a plan to (re)integrate both projects or will they be developed completly independant?

Re: openTSDB 1.1 vs openTSDB 2.0 ManOLamancha 2/26/13 7:49 AM
On Tuesday, February 26, 2013 5:38:16 AM UTC-5, bernd wrote:
So my question is:
Will openTSDB 2.0 include all upcoming features from openTSDB 1.1?
Is there a roadmap / a plan to (re)integrate both projects or will they be developed completly independant?

Actually the Annotations will be in 2.0, not 1.1. The 1.1 release just has a fix so that if someone tries 2.0, inserts some annotations, then downgrades to 1.1, it won't delete or error on the data in storage. Exactly what will turn out to be the official 2.0 is up in the air but be sure to tell us any features you like or want to see. Thanks
Re: openTSDB 1.1 vs openTSDB 2.0 tsuna 2/26/13 11:05 PM
I think the confusion stems from the fact that Brian decided to name
his fork "OpenTSDB2", but it's really another project, not OpenTSDB
anymore, because too many things have changed at once.  I asked him
kindly to rename his project to something else to avoid this sort of
confusion.

--
Benoit "tsuna" Sigoure
Re: openTSDB 1.1 vs openTSDB 2.0 bernd 2/27/13 12:53 AM

Am Mittwoch, 27. Februar 2013 08:05:36 UTC+1 schrieb tsuna:
On Tue, Feb 26, 2013 at 7:49 AM, ManOLamancha <clars...@gmail.com> wrote:
> On Tuesday, February 26, 2013 5:38:16 AM UTC-5, bernd wrote:
>>
>> So my question is:
>> Will openTSDB 2.0 include all upcoming features from openTSDB 1.1?
>> Is there a roadmap / a plan to (re)integrate both projects or will they be
>> developed completly independant?
>

I think the confusion stems from the fact that Brian decided to name
his fork "OpenTSDB2", but it's really another project, not OpenTSDB
anymore, because too many things have changed at once.  I asked him
kindly to rename his project to something else to avoid this sort of
confusion.

--
Benoit "tsuna" Sigoure

Hi tsuna,
you are right with your assumption,
I was confused and thougth that this http://code.google.com/p/opentsdb2/ is the"real" openTSDB.

Nevertheless there are some interessting features in http://code.google.com/p/opentsdb2/ that looks really nice, e.g. tree and the json interface.

Especialy the tree feature looks really nice for me, because users of SCADA Systems are accustomed too arrange plants in tree like structures.
The endpoints of SCADA Trees/Branches are sensors that produce time series which should be stored in an archvie application which uses openTSDB as backend.

It seems likely to use the SCADA tree path and store it as tag values like openTSDB 2.0 it does and use this information for tree visualisation in the archice application where each timeseries come from.

For me it would be really valuable when i could see how openTSDB develops in the near future.
What features will be implemented (maybe trees or json interface)?
Does a rough roadmap exist for the upcoming development?

Re: openTSDB 1.1 vs openTSDB 2.0 Paul-Armand Verhaegen 3/7/13 1:03 AM

Brian has changed the opentsdb2 to KairosDB, see http://code.google.com/p/kairosdb

Side note: However, because consistency is the last resort of the unimaginative, the github page for KariosDB is github.com/proofpoint/opentsdb2
And on that github page KairosDB is currently named KairosDB2, but I presume this is a state of flux.
Re: openTSDB 1.1 vs openTSDB 2.0 Brian Hawkins 3/7/13 7:40 AM
Ah man, you ruined the surprise.  Sheesh :(

It is still in flux.  Should be finished up in a day or two.

Brian
Re: openTSDB 1.1 vs openTSDB 2.0 Paul-Armand Verhaegen 3/12/13 6:23 AM

Yeah, Santa hates me too ;)  No, seriously, if I would have known, I would have kept the surprise.

Anyway, I do like where you're going with KairosDB, that's exactly the reason why I was watching over your shoulder.
And hopefully, my work will permit me to get involved a little bit more than just watching soon.
Re: openTSDB 1.1 vs openTSDB 2.0 Brian Hawkins 3/12/13 3:31 PM
Well we have officially done it now.  The new home of the formerly known as opentsdb2 project is https://code.google.com/p/kairosdb.  The new alpha release is sporting all kids of cool features such as:
  1. Stackable aggregators.  Sum over an hour and then average over a day with your data.
  2. Customizable aggregators.  Write your own aggregators to extend and customize KairosDB.
  3. Smaller cassandra schema.  Saves almost 50% of the space over the previous schema.
  4. Java client library.
  5. Packaged linux installers.  (RPM, DEB)
  6. Start and stop service scripts.
previously added feature but worth mentioning again:
  1. Millisecond time stamps (when using cassandra)
  2. HTTP rest api.
  3. Clean datastore abstraction.  Writing your own datastore is simple and straight forward.
  4. Compatible with tcollectors.
  5. Customizable UI.  Jetty serves up the UI from the webroot folder that you can extend.
  6. Built in H2 datastore for faster development work.
  7. Import / Export features.
  8. Works on both Cassandra and HBase (HBase support is provided by code compiled from opentsdb project to maintain compatibility)
  9. Client rendered UI
Thank for everyone's input and help.  To help minimize confusion between the two projects we have started our own discussion group at https://groups.google.com/forum/?fromgroups#!forum/kairosdb-group

Brian