Brian Hawkins

We have had more requests for this kind of thing. We will look into adding a group to the api. Also a somewhat related item to this issue is how...

I was messing around with this as I've hit the same problem viewing some collectd data. I created a branch feature/filterdiff where I've added a negative filter flag to the...

Oh yes, I like this idea.

I'm afraid that in 1.1.0 you were actually loosing data. If a buffer failed to send the retry never happened because Hector cleared the buffer. With 1.1.3 it will now...

This is an excellent point and I've addressed it in the CQL branch of the code. I'm implementing a file backed queue in Kairos so that it can buffer and...

I started that last comment this morning and just posted it, then I saw all your other remarks. I think the taking out of the exception was an oversight. I...

Thanks for doing all this, this is some good research. I assume you mean kairosdb.datastore.cassandra.write_buffer_max_size not write_delay.

The static mtime allows me to know when the last time a partition was updated, this is for the service key stuff. The frozen map is required in order to...

Thanks for the feedback. The real issue here is that writing data to Cassandra happens on a different thread from the incoming request. Data is buffered up in chunks and...

With the conversion to CQL I'm adding a queue in Kairos. Kairos will respond only after the data is in the file backed queue for durability.