1

I recently bought a managed Redis cluster on Digital Ocean, here are the specs: 1 GB RAM / 1vCPU / 10 GB Disk / Primary only.

I'm only using this cluster for a PUB-SUB system where an external application sends around 400 very small json messages to Redis and a client is listening for the messages. I thought that the load would be too much for the cluster, but instead Disk and memory usage are still fairly low.

Only thing worrying me is the fragmentation ratio, which is on average 3.8, but sometimes it spikes down to 1. Now how much should i be worried about this? I read that memory fragmentation tells how efficiently is memory allocated, so i got a little concerned about this. I'm pretty sure it depends on the loads of messages i send to Redis. Is there any way to improve this at all or can i only scale up my cluster? Any kind of advice is appreciated!

1 Answer 1

1

No need to worry about this at all. Fragmentation is quite meaningless when no/little RAM is used by Redis, as is in the case of pure Pub/Sub.

3
  • Thanks a lot! So i need to only keep track of RAM and CPU in this case?
    – JayK23
    Nov 29, 2020 at 9:47
  • Definitely worry about these, but also perhaps keep an eye on the network and the number of connections (file descriptors) the server maintains. Another potential monitoring spot is the SLOWLOG to detect long PUBLISH (due to immense subscribers lists). I think that's about ti. Nov 29, 2020 at 15:26
  • Thank you a lot again! As for the number of connections, there is only one publisher and one subscriber, both are python applicatio, so as for the number of connection that's it. According to Digital Ocean, CPU and Memory usage are at around 35 and 40%, which is good i think.
    – JayK23
    Nov 29, 2020 at 15:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.