New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Well, considering that the bandwidth limit is per bucket (not per object) and the pricing is based on 1TB increments it makes feel that it’s just a fancy wrapped single-node storage with S3 compatibility, rather than actual distributed object storage.
Good enough for Nextcloud, but wouldn’t put any throughput intensive stuff.
I haven't gotten any experience it with it, but looking at the Hetzner sub-reddit it seems like there are many complaints about degraded performance and periods of the service being down entirely. I don't know the severity of the issue or if it has improved but the initial reception there is generally poor.
While I generally am very happy with Hetzner their S3 offer just doesn't seem to competitive especially at this early stage.
LAYER7 FOREVER!!!!!
I've previously used it for ~3 months; pricing is good and clear, performance is also great.
Downsides: Lack of access key restriction controls, common service degradation notices, along with Nuremberg buckets always failed to be created for me (hangs and gets 403s).
Petty: Wish it was charged in smaller increments, maybe instead of 1TB it would be 100GB, maybe I'm just very frugal.
I have hopes for it, but I've since moved to BorgBase on the free tier. From my experience with S3, my own hosted MinIO in the past was most stable and most performant (and least redundant).
Do you think it’s good enough for Prometheus metrics data and Loki logs data?
Honestly @vitobotta , I have no experience with those two networking wise.
If I were in your position, since Hetzner doesn't currently charge for requests, I would reach out to their support and ask what the acceptable request rate is, as I know Prometheus would be based on 'how many metrics'*'how many nodes'.
For Loki logs, if you have the ability to tune the push rate, I cant see a issue arising.
Also Hetzner's Helsinki S3 seemed bizarrely more responsive (faster?), might be simply because it's being used less.
Thanks, I'll reach out to them then. We've been using Cloudflare R2 with Prometheus and Loki and I hadn't thought about request rate. Thanks
Depends on your log volume. Loki can easily pull a few gigabits per second if you search over large timeframes.
Exactly. It’s not even competitive price-wise, much like the direction Hetzner is going overall now
@Hetzner_OL
Do you have any tips on how to find out the amount of data transferred for long running queries?
Price-wise it's still a lot cheaper than Cloudflare R2 we use now. But the main reason is that since we are migrating from GCP to Hetzner, I thought that performance would be better if we had object storage in the same infrastructure.
It’s not even comparable to R2 IMO
You can either look at GCP's bucket monitoring or pod metrics. Kubelets have built-in cAdvisor that exposes network usage. It's available though Kubernetes API server -
/api/v1/nodes/${1}/proxy/metrics/cadvisor.Saw recently this
https://www.reddit.com/r/hetzner/comments/1jqeonu/too_many_issues_with_object_storage/
which layer7 product is the competing product?
Minio running on a LAYER7!!! VPS.
layer7 VPS has 200 IOPS and 75M/s limit on HDD, is that okay for an object storage?
@EthanZou - You tagged me above, but didn't really write a question or make a request. So I assume that you wanted me to see the feedback here so I can pass it onto my colleagues. I will do that.
We know Object Storage hasn’t been as stable as it should be, and we’re truly sorry for the hassle this has caused. We’re working hard to improve reliability and address the unforeseen challenges that have emerged from growing demand. We’re confident that we will be able to resolve these issues in the foreseeable future, and we'll continue to be a reliable partner for Object Storage.
Your feedback helps us along the way. So, let us know your concerns, and if you're facing major issues, please open a ticket so we can assist you directly. Thanks for sticking with us. -- Katie