Re: default_time_to_live setting in time series data

2018-01-11 Thread Eunsu Kim
Thanks for the quick response. TWCS is used.



> On 12 Jan 2018, at 11:38 AM, Jeff Jirsa  wrote:
> 
> Probably not in any measurable way. 
> 
> -- 
> Jeff Jirsa
> 
> 
>> On Jan 11, 2018, at 6:16 PM, Eunsu Kim  wrote:
>> 
>> Hi everyone
>> 
>> We are collecting monitoring data in excess of 100K TPS in Cassandra. 
>> 
>> All data is time series data and must have a TTL. 
>> 
>> Currently we have set default_time_to_live on the table.
>> 
>> Does this have a negative impact on Cassandra throughput performance?
>> 
>> Thank you in advance.
>> 
>> -
>> To unsubscribe, e-mail: user-unsubscr...@cassandra.apache.org
>> For additional commands, e-mail: user-h...@cassandra.apache.org
>> 
> 
> -
> To unsubscribe, e-mail: user-unsubscr...@cassandra.apache.org
> For additional commands, e-mail: user-h...@cassandra.apache.org
> 


-
To unsubscribe, e-mail: user-unsubscr...@cassandra.apache.org
For additional commands, e-mail: user-h...@cassandra.apache.org



Re: default_time_to_live setting in time series data

2018-01-11 Thread Jeff Jirsa
Probably not in any measurable way. 

-- 
Jeff Jirsa


> On Jan 11, 2018, at 6:16 PM, Eunsu Kim  wrote:
> 
> Hi everyone
> 
> We are collecting monitoring data in excess of 100K TPS in Cassandra. 
> 
> All data is time series data and must have a TTL. 
> 
> Currently we have set default_time_to_live on the table.
> 
> Does this have a negative impact on Cassandra throughput performance?
> 
> Thank you in advance.
> 
> -
> To unsubscribe, e-mail: user-unsubscr...@cassandra.apache.org
> For additional commands, e-mail: user-h...@cassandra.apache.org
> 

-
To unsubscribe, e-mail: user-unsubscr...@cassandra.apache.org
For additional commands, e-mail: user-h...@cassandra.apache.org



Re: default_time_to_live setting in time series data

2018-01-11 Thread kurt greaves
No. Nothing measurable. In fact it should be beneficial if paired with TWCS

On 12 Jan. 2018 1:17 pm, "Eunsu Kim"  wrote:

Hi everyone

We are collecting monitoring data in excess of 100K TPS in Cassandra.

All data is time series data and must have a TTL.

Currently we have set default_time_to_live on the table.

Does this have a negative impact on Cassandra throughput performance?

Thank you in advance.

-
To unsubscribe, e-mail: user-unsubscr...@cassandra.apache.org
For additional commands, e-mail: user-h...@cassandra.apache.org