All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
99.9% SLA on a low end VPS, is it possible?
vRozenSch00n
Member
I see quite a number of low end VPS providers proudly offer a 99.9% SLA.
99.9% uptime means the acceptable downtime slot is: ((100% - 99.9%) x (30dx24hx60m)) = 43.2 minutes/month or roughly 1.44 minutes/day.
That means should there be a problem related to the service, the problem should be resolved under 43.2 minutes.
If the resolution time takes 15 minutes each, then 3 downtime incidents in a month would breach the 99.9% SLA.
In this case a high availability system is needed as an effort to reduce or eliminate downtime, so when a system crashed, it should automatically fire up or switch to a life mirror and in case of the Internet connection disrupted, a backup connection should be available to be fired up at any time.
Unfortunately building such a high availability system is not cheap, and when a provider doesn't have much capital, the provider should have a sufficient amount of knowledge to maintain a cheapest possible high availability system.
I oversaw some cases in this forum related to downtimes that takes hours to be resolved caused by a hard drive failure, switch failure no backup and need to reprogram a new one, careless construction that cut a fiber optic connection that caused connection disruption, and many other problems that affect an entire node.
Those cases raised a simple question:
"Is a 99.9% SLA for real or is it only a marketing gimmick?"
What do you think?
Comments
Of course it's possible. We have 99.9% SLA for our VPS plans and honor it. Haven't been a problem so far.
Would you share how you do it, please?
It's simple - we own everything - the hardware, the network, the network hardware (routers, switches), etc. So we have full control over things. Also we have multiple upstreams, so one upstream dying doesn't leave us dead in the water.
I think so far we only had to pay SLA compensations once, when a node had some failure and was down for an hour and some minutes. We automatically applied the SLA credits to all affected customers.
@rds100 That's very good. What monitoring system do you use? Is it in-house built?
in-house. It's not perfect, but we had a lot of years to develop it
We are in the ISP business since around 1998.
That's impressive. So, do you have your hands dirty setting up critical things or do you hire someone else to do it?
We do everything. No external people / outsourcing.
Actually some other businesses outsource some things to us ;-)
So, that's the fruit of long time experience
@rds100 How do you plan and manage your customer, do you separate low resource user in one node and mid-high resource user in another?
Do you mix several package in one node or do you separate them in other nodes?
@vRozenSch00n we don't have any ultra-lowend plans. I don't like tons of VPSes on one node. Our plans are not the cheapest as you can see, so we can afford to put less VPSes per node and still be profitable.
I see, so this is also helps to maintain stability.
we have that amount sla on our vps plan
and if there are network outage more than 0.1 we will give 1 free month to our client
and if there are network outage more than 0.1 we will give 1 free month to our client
Thank you, @tuguhost. In case of network outage, suppose it is not on your side, does your upstream compensate you?
No
thats would be my risk
thats would be my risk
I see, so it is part of your good service to your customer, isn't it?
thats because i advertise that and my upline not
it's possible ...
Hi brother, it's good to see you grow and have a good range of offers at your site.
If you're housed in a tier 3 DC, then sure. If elsewhere, it may be possible but my confidence of you pulling it off drops. Not personal or anything, just being realistic with the known probabilities.
That's kind. Many LEB hosts send out RFO with "if you would like to claim SLA submit a support ticket.." which sound "if you really need to..." so I imagine that most clients don't even bother with SLA much.
4 out of 5 of my LowEndBoxes are either 99.9% or 100%. I'm very impressed with the LET/LEB market.
I can kind of understand why some providers ask you to open a ticket. When i applied the credits to all our affected clients it was a very hard manual job going through all the clients, checking if they were on the affected node, calculating the credits, etc. Probably the time spent doing all this was worth more than the total sum of credits applied
Sometimes host do some big changes and announces it. And then the big change causes delays leaving hours of downtime. Then since its planned, not part of SLA
We do provide That SLA,
similar to rds100 we own everything, have redundant power , have redundant uplink providers and all routers and switches are redundant.
and we add another layer to that topology by making even the servers run in high availability pool. so if one server fails the remaining servers will host the vms of that server.
Again we are not the cheapest here
Dunno if @INIZ has it but my vps have had 99%+ uptime close to every month for 14 months
I think a good majority of the top 3 providers probably fall into or close to 99% uptime. Personally my experience has been around there with at least Hostigation / Ramhost. I am on RamNode which had down time when I got it, but I believe it's been fine since and am pretty sure it will be up there with the other two since they seem to be very promising.
I considered picking up an INIZ too, but I remember reading something about drama some time ago with their previous company name, so instead of researching into it I just decided I have enough. :P Though I might pick up another Ramnode if it proves to have nice uptime as I believe it will based on all the community feedback/personal experience.
Well, cheap is a relative term. A $1/month can be expensive when there is too much problems and downtimes.