Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

How much Uptime Percentage is acceptable to be called as a good host?
New on LowEndTalk? Please Register and read our Community Rules.

How much Uptime Percentage is acceptable to be called as a good host?

Many providers advertise their servers with a 100% Uptime tag. And it is very hard to believe, as there can be many unforeseen issues like power outage, network down, router failure, disk failure and list goes on...

But anyway, as per your personal optional what percentage of uptime can be considered as acceptable. Every customer would expect 100% Uptime, but for now let us speak about the reality.

Comments

  • LTnigerLTniger Member

    Depends. For example paying 5 USD / quarter for 500 GB storage I expect at least 70% uptime. The rest 30% downtime is expected as server is used only to store cold backups.

    #!/Bashblog.net | Free Wordpress Hosting | If you can't idle, what's the point?

  • SirFoxySirFoxy Member

    99.9%

    Thanked by 1m4nu

    lurking in the shadows like a wombat or some shit

  • m4num4nu Member, Provider
    edited May 3

    99.90 to 99.95 is the economical sweet spot, I think. This would equal 4 to 8 hours of downtime per year. Anything better than 99.99 will require a more complex redundant setup. Will be hard to do with a single server or location.

    You could improve the "subjective" uptime by having a fixed maintenance window at an off-peak time. Works well if your users are geographically in the same place or have similar usage patterns.

    Simple and secure Borg Backup hosting from $2/month: BorgBase.com

  • tetechtetech Member

    @vladimirlenin said:
    Many providers advertise their servers with a 100% Uptime tag. And it is very hard to believe, as there can be many unforeseen issues like power outage, network down, router failure, disk failure and list goes on...

    But anyway, as per your personal optional what percentage of uptime can be considered as acceptable. Every customer would expect 100% Uptime, but for now let us speak about the reality.

    A X% uptime guarantee does not imply the provider expects to achieve X% uptime, just that they will issue some type of credit (according to their SLA) if the uptime is less than X%.

    It is similar to a manufacturer offering a warranty on the car for 3 years - it does not mean they expect no car will have a problem for 3 years, just that they will cover the repairs when something goes wrong during that time.

    Second, many of the "uptime" guarantees I've seen exclude issues with upstream providers which limits the usefulness.

    Personally I'd put any host under 99.95% on a watch list and consider them for replacement. But I also look at the pattern of outages. I've got one host that is at 99.78% which I'm keeping (for now) since the problem was one solid block where a router failed. I find that much easier to work around than providers which have higher average uptime but where I'm getting shorter duration outages multiple times per week.

    Thanked by 3yoursunny zxrlha Ouji
  • @vladimirlenin said:
    Many providers advertise their servers with a 100% Uptime tag. And it is very hard to believe, as there can be many unforeseen issues like power outage, network down, router failure, disk failure and list goes on...

    But anyway, as per your personal optional what percentage of uptime can be considered as acceptable. Every customer would expect 100% Uptime, but for now let us speak about the reality.

    @m4nu said:
    99.90 to 99.95 is the economical sweet spot, I think. This would equal 4 to 8 hours of downtime per year. Anything better than 99.99 will require a more complex redundant setup. Will be hard to do with a single server or location.

    You could improve the "subjective" uptime by having a fixed maintenance window at an off-peak time. Works well if your users are geographically in the same place or have similar usage patterns.

    I would narrow it more from 99.95 to 99.99

  • skorupionskorupion Member
    edited May 3

    99.997% excl. maintenance that was advertised at least 2 weeks in advance.

    My Blog rafalblog.xyz
    Contabo GmbH insanely cheap VPS |

  • deankdeank Member, Troll
    edited May 3

    100% uptime.

    Any sort of downtime should result in a SLA credit of 12 years of free service.

    No maintenance allowed.

    "Jarland is stupid."

  • sayem314sayem314 Member

    99.995%

    Downtime

    • year: 26.30 minutes
    • month: 2.19 minutes

    To The Moon!

  • zxrlhazxrlha Member

    99.9% uptime and less than one down period per month is minimal requirement, and if other aspects(performance,price,etc) are good then I think it is a good host.

    If other aspects are normal/average, at least 99.99% uptime is needed to be called as a good host.

    QuantumCore :) Webhorizon :) LittleCreek :D Clouveo ;) RackNerd ;) NexusBytes :* Novos :p NaranjaTech :* QuickPacket :blush:

  • momkinmomkin Member

    @deank said:
    100% uptime.

    Any sort of downtime should result in a SLA credit of 12 years of free service.

    No maintenance allowed.

    Your website link please sir, i want to order from you sir.

  • dahartigandahartigan Member without signature

    I think 99.99% is a good target and all, but realistically 98-99% works for me and my projects.

  • deankdeank Member, Troll

    @momkin said:
    Your website link please sir, i want to order from you sir.

    https://endis.nigh/order.php?referral+ID=xxxdeank666xxx

    "Jarland is stupid."

  • ben47955ben47955 Member

    Uptime is not the only factor to call a good host or a bad host.

  • momkinmomkin Member

    @deank said:

    @momkin said:
    Your website link please sir, i want to order from you sir.

    https://endis.nigh/order.php?referral+ID=xxxdeank666xxx

    Not working , or is it too early for the summer ?

  • deankdeank Member, Troll

    It means you are not a member of Nigh cult.

    "Jarland is stupid."

  • 97% up is very good..

    A simple uptime dashboard using UptimeRobot API https://upy.duo.ovh
    Currently using VPS from BuyVM, Gullo's, Hetzner, HostHatch, HostSailor, HostSolutions, InceptionHosting, LetBox, MaxKVM, MrVM, VirMach.

  • tetechtetech Member

    @chocolateshirt said:
    97% up is very good..

    :-0

    Maybe for some type of idler or backup storage that is written to for a few minutes per day. For anything actually being used, if it goes below 99.5% I am usually taking it out of service and writing off the remaining prepaid $.

  • GameTownProjectsGameTownProjects Member, Provider

    I think around 99÷ should be fine for the most users, BUT Microsoft, wich is by far not the cheapest, offers only 95÷ on HDD instances.

    I think the most important thing is that the provider knows how much he can promise, however, unfortunately, very many specify 100% or similar and then can not keep it.

    We state our availability at 99.5%.

  • msallak1msallak1 Member
    edited May 4

    Arubacloud are saying their Czechia Data Center has 100% since 2004! Could that even be possible?
    https://www.arubacloud.com/infrastructures/czech-republic-dc-cz1.aspx
    Did I miss something here?

    Edit: Corrected year

  • PulsedMediaPulsedMedia Member, Provider

    @msallak1 said:
    Arubacloud are saying their Czechia Data Center has 100% since 2004! Could that even be possible?
    https://www.arubacloud.com/infrastructures/czech-republic-dc-cz1.aspx
    Did I miss something here?

    Edit: Corrected year

    Depends how you measure it. That the building has been standing for that long? Sure not a problem.

    That long without complete power outage, always at least something is powered in the building? Sure not a problem.

    It is all about what and how you measure :)

    Also it's about luck and statistics, you can be lucky and have with no redundancy what so ever 100% uptime for that time down to a single server; But that is extremely extremely unlikely.

  • james50ajames50a Member

    Honestly as long as any downtime is either communicated in advanced ie hardware upgrades or from stuff outside of providers control ie flood/fire id say its good enough.

  • nurexnurex Member

    For me I don't like numbers like 99.9%, but I have two requirements:

    • Don't oversell a loooooot so machine goes down several times a day;
    • If there's problem, just work to fix it.

    If a provider can't even achieve these, then for me it's time to leave.

  • @nurex said:

    • Don't oversell a loooooot so machine goes down several times a day;

    But how can you find out whether it is oversold or not???

  • @vladimirlenin said:

    @nurex said:

    • Don't oversell a loooooot so machine goes down several times a day;

    But how can you find out whether it is oversold or not???

    If performance is dog shit - the host is likely overselling.

  • chipchip Member

    @vladimirlenin said:
    Many providers advertise their servers with a 100% Uptime tag. And it is very hard to believe, as there can be many unforeseen issues like power outage, network down, router failure, disk failure and list goes on...

    But anyway, as per your personal optional what percentage of uptime can be considered as acceptable. Every customer would expect 100% Uptime, but for now let us speak about the reality.

    As close to 100% as humanly possible... stuff happens and things break but how quickly can you/do you fix them?

    100% is a sales gimmick, but 99.99% should be achievable most of the time (that gives something like 30-60 mins downtime a month) and if things are breaking every month something needs replacing

  • HassanHassan Member, Provider

    @chip said:

    @vladimirlenin said:
    Many providers advertise their servers with a 100% Uptime tag. And it is very hard to believe, as there can be many unforeseen issues like power outage, network down, router failure, disk failure and list goes on...

    But anyway, as per your personal optional what percentage of uptime can be considered as acceptable. Every customer would expect 100% Uptime, but for now let us speak about the reality.

    As close to 100% as humanly possible... stuff happens and things break but how quickly can you/do you fix them?

    100% is a sales gimmick, but 99.99% should be achievable most of the time (that gives something like 30-60 mins downtime a month) and if things are breaking every month something needs replacing

    this is the correct answer

  • OhJohnOhJohn Member

    99.99% is less then 5min a months. 99.93% would be about 30min. So it sounds as if you meant 99.9%, not 99.99%.

  • AXYZEAXYZE Member

    @chip said:

    @vladimirlenin said:
    Many providers advertise their servers with a 100% Uptime tag. And it is very hard to believe, as there can be many unforeseen issues like power outage, network down, router failure, disk failure and list goes on...

    But anyway, as per your personal optional what percentage of uptime can be considered as acceptable. Every customer would expect 100% Uptime, but for now let us speak about the reality.

    As close to 100% as humanly possible... stuff happens and things break but how quickly can you/do you fix them?

    100% is a sales gimmick, but 99.99% should be achievable most of the time (that gives something like 30-60 mins downtime a month) and if things are breaking every month something needs replacing

    You mean 99.9% right?

    Because with 99.99% it would be 4m 22s per month
    With 99.9% its 43m 49s

  • jarjar Provider

    1% per USD

    Thanked by 1SmallWeb

    Founder @ MXroute

  • SmallWebSmallWeb Member, Provider

    @jar said:
    1% per USD

    7% Debian thx

    Thanked by 1jar

    Michael from SmallWeb - Please use official support methods for help.

  • chipchip Member

    @AXYZE said:

    @chip said:

    @vladimirlenin said:
    Many providers advertise their servers with a 100% Uptime tag. And it is very hard to believe, as there can be many unforeseen issues like power outage, network down, router failure, disk failure and list goes on...

    But anyway, as per your personal optional what percentage of uptime can be considered as acceptable. Every customer would expect 100% Uptime, but for now let us speak about the reality.

    As close to 100% as humanly possible... stuff happens and things break but how quickly can you/do you fix them?

    100% is a sales gimmick, but 99.99% should be achievable most of the time (that gives something like 30-60 mins downtime a month) and if things are breaking every month something needs replacing

    You mean 99.9% right?

    Because with 99.99% it would be 4m 22s per month
    With 99.9% its 43m 49s

    Yes 99.9%

  • PandoGulfPandoGulf Member

    100% when needed. 0% if it's idled.

  • jsgjsg Member
    edited May 5

    Some seem to have forgotten where they are. Hint: at low end talk. Plus, quite a few seem to not know how consistently high availability is achieved; that needs quite a bit of both knowledge and experience as well as money.

    My expectation for anything up to $5/mo is 99.95% and I'm glad to see that quite a few providers here achieve that or even do even better.
    Between $5 and $10 per month I expect 99.99% and I'm amazed that quite a few providers here do deliver that at about $7/mo.

    For professional use I expect a 99.995% SLA (and no "maintenance window" BS) and if broken even a single time I'll leave - but there $20 or even $50 per month is no problem.

    Plus I expect an announcement no less than a week in advance for maintenance downtime or else, in the case of an emergency, an immediate public statement and info.

    That said, if uptime/availability is really critical we ourselves must do our part too, e.g. by creating/using shadowing, mirroring, etc. our critical applications/services.

    The problem with democracy is that by definition > 85% of the voters are not particularly intelligent.

  • donkodonko Member

    High as possible without any network maintenance, node reboot, migrations and more random issues every week.

  • pierrepierre Member

    @donko said:
    High as possible without any network maintenance, node reboot, migrations and more random issues every week.

    @cociu take some notes^

  • jsgjsg Member

    ... and fried quails flying into our mouths of course ...

    Ask for something reasonably achievable and (at least some) providers will try to achieve it, but ask for wet dreams and all you'll get is a bad awakening.

    The problem with democracy is that by definition > 85% of the voters are not particularly intelligent.

  • asasdasasd Member

    4 reasons why the "100% uptime guarantee" phrase falls in the false/misleading advertisement category:

    1. Vis mayor events are NEVER included in the guaranteed uptime percentage.
    2. If the provider doesn't meet the SLA, the compensation will likely not be automatic, and the amount will be negligible. That's why trading companies do build their own networks.
    3. The advertised uptime does not include planned maintenance (shoutout to the very few exceptions whom I can count on one hand). Even the real high-availailability providers with own cables, multi-datacenters and disaster recovery plans define service windows.
    4. Uptime history of recent years is not included — usually you won't find any links to independent 3rd-party monitoring services either.

    What percentage is acceptable?

    It depends on the reason of outages. If there were no major kernel/software/hardware updates during the month, I think 99.95% (21 minutes and 55 seconds monthly downtime including planned maintenance) should be the minimum. That percentage is easily reachable even with daily reboots.

Sign In or Register to comment.