Do you have the SSTables per read metric for before and after you increased
the key cache size? If it was high before, that may have been the culprit
meaning compaction tuning is in order.
On Fri, Aug 25, 2023, 12:35 PM Shaurya Gupta wrote:
> Thanks everyone.
> Updating this thread -
> We increa
For a readiness probe and for ongoing ECV checks, just making sure the
driver is initialized is enough. I've seen problems recently with
applications running "select cluster_name from system.local" for ECV
checks. We haven't dug into it in detail yet but with a large number of
clients it puts a lo
That's a good way to it!
On 25/08/2023 20:10, Shaurya Gupta wrote:
> We don't plan to open a new connection. It should use the same connection(s)
> which the application will use.
>
> On Fri, Aug 25, 2023 at 10:59 AM Raphael Mazelier wrote:
>
>> Mind that a new connection is really costly for C
We don't plan to open a new connection. It should use the same
connection(s) which the application will use.
On Fri, Aug 25, 2023 at 10:59 AM Raphael Mazelier wrote:
> Mind that a new connection is really costly for C*.
> So at startup it's fine. but not in a liveness or readiness check imo.
>
“select * from …” without a predicate from a user table would be very
expensive, yes.
A query from a small, node-local system table such as “select * from
system.peers” would make a better health check. 👍
- Scott
> On Aug 25, 2023, at 10:58 AM, Raphael Mazelier wrote:
>
>
> Mind that a new
Mind that a new connection is really costly for C*.
So at startup it's fine. but not in a liveness or readiness check imo. For the
query why not select 1; ?
--
Raphael Mazelier
On 25/08/2023 19:38, Shaurya Gupta wrote:
> Hi community
>
> We want to validate cassandra connectivity from the appl
Hi community
We want to validate cassandra connectivity from the application container
when it starts up and before it reports as healthy to K8s. Is doing
> select * from our_keyspace.table limit 1
fine Or is it an inefficient query and should not be fired on a prod
cluster ?
Any other suggesti
Thanks everyone.
Updating this thread -
We increased the key cache size from 100 MB to 200 MB and we believe that
has brought down the latency from 40 ms p95 to 6 ms p95. I think there is
still scope for improvement as both writes and reads are presently at p95 6
ms. I would expect writes to be low
