Page 1 of 2 12 LastLast
Results 1 to 40 of 44
  1. #1
    Join Date
    Feb 2006
    Location
    New York, NY
    Posts
    92

    How much downtime is too much downtime?

    So I'm always one that gives my hosts the benefit of the doubt, but I was just curious to see from others how much downtime from one host will make you pack up your bags and leave?

  2. #2
    Join Date
    Aug 2006
    Location
    London
    Posts
    549
    Any unplanned downtime is bad downtime, if it occurs and your datacenter doesn't give you an acceptable explanation - I'd start packing up.
    Ekin Ersoy
    Stagnom Servers | Europe Dedicated Servers with Redefined Support
    Follow us on Twitter: @StagnomUK | Like us on Facebook: facebook.com/Stagnom

  3. #3
    Join Date
    Mar 2009
    Location
    Santa Monica, CA
    Posts
    3,372
    Interesting, no one has considered the time-frame involved.

    For matters under their control, a shared hosting environment should easily provide over 95% each month (that's 36 hrs/month downtime). If they can't do that you've got some incompetents working for you

  4. #4
    Join Date
    Feb 2005
    Location
    Australia
    Posts
    5,842
    I expect a quality shared host to provide around 99.9% average, over a long period - say 6 months or more. In order to do that they'll need to achieve 99.95%-100% most months because very rare problems will occur (perhaps once a year or so) causing several hours of downtime. And I make no allowance for scheduled maintenance in this - it's better to be informed of downtime in advance but in the end the server's either working or it's not.

    The hosts that can achieve that sort of uptime tend to be the ones offering more expensive, smaller plans. For a budget plan with generous disk and transfer limits I'd adjust my expectations down to around 99.5% average. And any host who delivers less than 98% over a single month (14 hours downtime) needs a good explanation and a convincing case for why it won't happen again, or I'll be leaving.

    Just be careful of fake uptime "guarantees" and worthless SLAs. It really makes no difference what the host claims to offer - proven performance is what matters.
    Chris

    "Some problems are so complex that you have to be highly intelligent and well informed just to be undecided about them." - Laurence J. Peter

  5. #5
    Join Date
    Sep 2007
    Posts
    1,016
    Quote Originally Posted by foobic View Post
    Just be careful of fake uptime "guarantees" and worthless SLAs. It really makes no difference what the host claims to offer - proven performance is what matters.
    I would have to agree with you, however would you stick with a host who actually honors their SLA or uptime guarantee?

  6. #6
    Join Date
    Nov 2007
    Location
    Photoshop
    Posts
    382
    We don't provide an uptime guarantee although our upstreams give us 100% guarantees because lets face it servers will go down at some point or another.

    Maintenance is also sommet we have to factor in - like other hosts we do all we can to ensure we don't have to take sites offline. For example our servers operate hot-swap raid, dual nic's etc

    Sometimes its not avoidable though but I would say between 30 - 120mins of downtime every few months is acceptable (I have very high standards).
    Rob - Creative Director
    Peartree Digital Media Solutions & Peartree Hosting Department.
    Providing a range of services from graphic design and XHTML coding to web hosting, VPS and dedicated servers.
    My views and opinions are my own and in no way do they represent the company I work for...

  7. #7
    Join Date
    Feb 2005
    Location
    Australia
    Posts
    5,842
    Quote Originally Posted by InstantPH View Post
    would you stick with a host who actually honors their SLA or uptime guarantee?
    A guarantee without an SLA is meaningless (and arguably false advertising). SLAs are almost universally worthless. If you suffer 24 hours downtime in a month and your host (following SLA) gives you $10 credit, will that make you happy to stay?

    For me the SLA is irrelevant - the host's explanation of what happened and their plans to prevent it happening again would be the main things that influence the stay / go decision.
    Chris

    "Some problems are so complex that you have to be highly intelligent and well informed just to be undecided about them." - Laurence J. Peter

  8. #8
    Join Date
    Apr 2009
    Posts
    835
    48 hours in a row is definitely too much.

  9. #9
    Join Date
    Feb 2008
    Location
    Houston, Texas, USA
    Posts
    2,955
    How much downtime have you experienced? Any facility will have down time (unplanned). We've long accepted that failover across facilities must be part of our business continuity plan.

    Regards
    UNIXy - Fully Managed Servers and Clusters - Established in 2006
    [ cPanel Varnish Nginx Plugin ] - Enhance LiteSpeed and Apache Performance
    www.unixy.net - Los Angeles | Houston | Atlanta | Rotterdam
    Love to help pro bono (time permitting). joe > unixy.net

  10. #10
    Quote Originally Posted by Collabora View Post
    Interesting, no one has considered the time-frame involved.

    For matters under their control, a shared hosting environment should easily provide over 95% each month (that's 36 hrs/month downtime). If they can't do that you've got some incompetents working for you
    Holy Crap. I think I'd bounce way before 95% uptime lol. Anything under 99.5% in my opinion is totally unacceptable, even for shared hosting.

  11. #11
    Join Date
    Jun 2006
    Location
    NYC
    Posts
    1,446
    If it drops below 99% there is an issue.
    FiberPeer.Com | | REAL DDoS Protection | Cloud Hosting | VPS | Dedicated Servers | High Bandwidth Hosting | 1Gbps-10Gbps Unmetered
    FiberPeer DDoS Mitigation | ethProxy Upgraded! | 14-Years Experience | Emergency 24/7 Support
    Visit us @ www.fiberpeer.com

  12. #12
    Join Date
    Mar 2009
    Location
    Miami, Florida
    Posts
    19,167
    Hello,
    Here is my take on Uptime. Everyone must remember that there are different reasons for downtime; planned maintenance, updates, server reboots, unplanned maintance, abuse, DDoS or DataCenter network issues.

    A host should do the following
    • ALWAYS inform their customers of planned maintenance, even if it only has a remote chance of client impact. It is best to e-mail all clients at least 48 hours in advanced if known.
    • Have a network news page.
    • Strive for 100% Uptime
    Keith I Myers
    KMyers.me The rantings of a lunatic
    Geek Survival Guide - Reviews and Advice for Geeks

  13. #13
    Join Date
    Aug 2008
    Location
    Vancouver, Canada
    Posts
    650
    In my personal opinion any downtime longer than 12 hours is way too much. A host should try to keep downtime to as little as possible (hopefully 0 hours).
    Tailored VPS offers fully customizable VPS Hosting
    Powered by OpenVZ | Servers located in the USA | 99.9% Uptime

  14. #14
    Join Date
    Mar 2009
    Location
    Santa Monica, CA
    Posts
    3,372
    Quote Originally Posted by AquariusADMIN View Post
    Holy Crap. I think I'd bounce way before 95% uptime lol. Anything under 99.5% in my opinion is totally unacceptable, even for shared hosting.
    That's 7.2 minutes in a day. You would really change hosts if your site was down between 11:00 pm and 11:08 pm. You are the one that's crazy!

  15. #15
    Join Date
    Jul 2005
    Location
    Edinburgh
    Posts
    3,878
    Quote Originally Posted by Collabora View Post
    That's 7.2 minutes in a day. You would really change hosts if your site was down between 11:00 pm and 11:08 pm.
    Would I bother? no probably not....but most likely my clients who don't fully understand hosting etc would and as they're the ones who help pay my food and energy bills and my mortgage, then I'd need to protect their interests online.

    I get phone calls from a few clients even if I reboot the server they're on lol. Some clients must live on their webpages.

    I agree 8 minutes a day doesn't sound long....however, try hanging yourself by the bollocks for a few minutes and you'll soon realise even 2 minutes is a long time mate.

    owm
    Last edited by Outlaw Web Master; 07-27-2009 at 01:53 AM.
    ‹(¿)›
    Life's what you make it.

  16. #16
    Join Date
    Aug 2006
    Location
    London
    Posts
    549
    Quote Originally Posted by Outlaw Web Master View Post
    I agree 8 minutes a day doesn't sound long....however, try hanging yourself by the bollocks for a few minutes and you'll soon realise even 2 minutes is a long time mate.
    Awesome post
    Ekin Ersoy
    Stagnom Servers | Europe Dedicated Servers with Redefined Support
    Follow us on Twitter: @StagnomUK | Like us on Facebook: facebook.com/Stagnom

  17. #17
    It might be 10 mins of the downtime and that is the google update which is very important to you. I suppose in this case that would be to much for the downtime.

  18. #18
    Join Date
    Jul 2009
    Location
    Sputnik's World
    Posts
    82
    Quote Originally Posted by Collabora View Post
    That's 7.2 minutes in a day. You would really change hosts if your site was down between 11:00 pm and 11:08 pm. You are the one that's crazy!
    Downtime every single day? Doesn't really matter if it's only 7 minutes, if they can't keep the server up for more than 24 hours without it going down, there is a problem, and if they aren't willing to figure out what it is and fix it, then it's time to find a new host.

    As for the rest, I guess it depends on how frequent it is, how long it's down when it is down, and how badly it annoys me while it's down.

    I'm always willing to give them the benefit of the doubt, and I can be understanding about downtime, it does happen. Hardware dies, lines get cut during construction season, hurricanes happen along the coasts, earthquakes. I expect them to have backup plans for when it does happen, but I know those things will result in at least SOME downtime.

    It shouldn't however, be happening all the time, if it is, there is something wrong and it needs to be fixed.

  19. #19
    Downtime every single day?
    If that is the truth I suppose you have nothing to do about it but move. Better find another host that be waiting for the next downtime

  20. #20
    Join Date
    Jun 2007
    Location
    GA, USA
    Posts
    77
    I run low traffic sites so I'm currently not too obsessed with perfection, but if it's going to be less than 98% I'd leave. Exception is of course planned maintenance, and if it's planned it should still be as short as possible downtime. Ex: When my webhost moved to a new datacenter, they were able to somehow keep servers offline/online for about 5 hours (confuses me; I think they ghosted HDDs and were able to switch from one data center to the other), and that was cool with me. They have weeks of notice + limited downtime for what they were doing.

  21. #21
    Join Date
    Feb 2002
    Posts
    1,137
    Quote Originally Posted by Collabora View Post
    Interesting, no one has considered the time-frame involved.

    For matters under their control, a shared hosting environment should easily provide over 95% each month (that's 36 hrs/month downtime). If they can't do that you've got some incompetents working for you
    That's way too much downtime to be having each month, any host with 5% dowtime each month (or even 1% really) should sell up and do something else.

  22. #22
    Join Date
    Nov 2007
    Location
    Photoshop
    Posts
    382
    Presumably emergency maintenance and planned maintenance are excluded?
    Rob - Creative Director
    Peartree Digital Media Solutions & Peartree Hosting Department.
    Providing a range of services from graphic design and XHTML coding to web hosting, VPS and dedicated servers.
    My views and opinions are my own and in no way do they represent the company I work for...

  23. #23
    Join Date
    Jun 2009
    Location
    Singapore
    Posts
    202
    I have quite high standards. Anything below 99.8% uptime is unacceptable, although 99.9% is still best.

    Lets face it - we have to reboot our servers from time to time.
    bikster.com - Quality Hosting. Affordable Prices.
    Providing premium quality shared and reseller cPanel/WHM hosting at low prices!
    Reseller cPanel/WHM hosting solutions that you can afford

  24. #24
    Join Date
    Oct 2002
    Location
    EU - east side
    Posts
    21,913
    That's 7.2 minutes in a day. You would really change hosts if your site was down between 11:00 pm and 11:08 pm. You are the one that's crazy!
    ~4 hours of downtime per month may not be too much for most sites, I can agree to that. But it doesn't mean it's the acceptable level. When 99.9%+ yearly uptime hosting can be had for a small or even no increase in costs, why not take it?

    I'm with foobic on this issue. Real uptime is what matters. I don't care that you have to repair your generators every 2 days and that the downtime is thus "scheduled". I don't care for SLAs and guarantees because the compensation is usually not worth the time it takes to request it, not to mention the aggravation caused by the downtime itself.

  25. #25
    Quote Originally Posted by Outlaw Web Master View Post
    Would I bother? no probably not....but most likely my clients who don't fully understand hosting etc would and as they're the ones who help pay my food and energy bills and my mortgage, then I'd need to protect their interests online.

    I get phone calls from a few clients even if I reboot the server they're on lol. Some clients must live on their webpages.

    I agree 8 minutes a day doesn't sound long....however, try hanging yourself by the bollocks for a few minutes and you'll soon realise even 2 minutes is a long time mate.

    owm
    I know exactly what you mean clients eh - they can be a nightmare!

  26. #26
    Join Date
    Jul 2009
    Posts
    38
    I would prefer no more than 99.9% because it is very troublesome if it occurs longer than that.

  27. #27
    Join Date
    Apr 2009
    Location
    EU
    Posts
    602
    Quote Originally Posted by Collabora View Post
    Interesting, no one has considered the time-frame involved.

    For matters under their control, a shared hosting environment should easily provide over 95% each month (that's 36 hrs/month downtime). If they can't do that you've got some incompetents working for you
    Inno would be giving us 100% credit on 95% uptime.

    99.5 - 99.89 = 70% service credit
    99.0 - 99.49 = 80% service credit
    98 - 99 = 90% service credit
    97.99 - 95.00 = 100% service credit
    ______________________________________________________
    HostYourIdea - Serving you for 12 years

  28. #28
    Join Date
    Mar 2009
    Location
    Santa Monica, CA
    Posts
    3,372
    Quote Originally Posted by HostYourIdea View Post
    Inno would be giving us 100% credit on 95% uptime.

    99.5 - 99.89 = 70% service credit
    99.0 - 99.49 = 80% service credit
    98 - 99 = 90% service credit
    97.99 - 95.00 = 100% service credit
    But that is a single incident, web server only. "Customers claiming credit under our uptime guarantee must do so by opening a ticket within 72 hours of the incident " If downtime is not in a single block of time how do you make a claim?

  29. #29
    Join Date
    Jul 2009
    Location
    Sputnik's World
    Posts
    82
    Quote Originally Posted by Rob View Post
    Presumably emergency maintenance and planned maintenance are excluded?
    That depends.

    World of Warcraft has planned maintenance every week, it runs about 6 hours on Tuesday. Would that be acceptable "planned" maintenance for your webhost? Would you go with a host if you knew that was their weekly planned maintenance? (We won't go into the downtime for their last hardware upgrade.)

    The answer to your question really depends on the length and frequency of both the regular maintenance and the emergency maintenance.

  30. #30
    Join Date
    Mar 2009
    Location
    Santa Monica, CA
    Posts
    3,372
    A lot of people here are disquising "downtime" with "planned maintenance." Anything can be planned maintenance if the downtime is planned anytime in future, e.g., "In 3 hours at 10pm EST we are taking servers down for emergency maintenance." Translation: We screwed something up and now we have to fix it

  31. #31
    Join Date
    Apr 2009
    Location
    EU
    Posts
    602
    Quote Originally Posted by Collabora View Post
    ...If downtime is not in a single block of time how do you make a claim?
    Well I was only thinking of one event. At least you get a free month if they do go down.

    If they are below 99% on average, then it's time to pack your bags.
    ______________________________________________________
    HostYourIdea - Serving you for 12 years

  32. #32
    Join Date
    Mar 2009
    Location
    LAX, DAL, MIA, ATL, ORD
    Posts
    8,955
    Quote Originally Posted by serverorigin View Post
    If it drops below 99% there is an issue.
    I'd say the same thing. Also as i tell everyone do your own back up's just incase you have to move to a new host and your old host has been offline and you cant retrieve the back up because they have been offline.

  33. #33
    Join Date
    Mar 2009
    Location
    Santa Monica, CA
    Posts
    3,372
    Quote Originally Posted by HostYourIdea View Post
    Well I was only thinking of one event. At least you get a free month if they do go down.

    If they are below 99% on average, then it's time to pack your bags.
    Still, this is a useless concept. What do you mean by down? What if you can ping the server but the customer's page does not load for some reason? Is that up or down? Do you ping server to measure downtime over time? If so, what is the ping interval? Its easy to get near 100% uptime if its defined by simply by ping response and you ping every 1 to 10 minutes. For example, if you use ping to measure downtime and your down half the time you can still be showing 100% uptime, if all the downtime is between pings, but your actual uptime is really only 50%. In this regard, I agree that you should be able to pull off near 100% uptime according to ping reports.

    <I am using the word "you" in the editorial sense>
    Last edited by Collabora; 07-27-2009 at 03:05 PM.

  34. #34
    Join Date
    Oct 2008
    Posts
    533
    I don't really mind the 1-2 minute downtimes for the occasional server reboots since I don't run a business orientated site. But frequent downtimes spanning over 15 minutes will get me searching on the shared offers section.

  35. #35
    Crickey, id have to start tunneling out of our office and re-enforcing doors if our servers went down for 10 minutes every day. Somehow (despite the servers being monitored by Nagios every few seconds) our customers still manage to somtimes submit a ticket when a web server (apache) has died before nagios picks it up. I often wonder how they notice soo incredibly quickly!

    Would you guys give a host more of a chance if they kept you up to date of the situation on a blow by blow basis but had a bit more downtime say 99% compared to a host who gives little information other than "we are down, looking at it" but has a 99.5% uptime record?

    Am just curious

  36. #36
    Join Date
    Mar 2009
    Location
    Santa Monica, CA
    Posts
    3,372
    Quote Originally Posted by andrewjab View Post
    Would you guys give a host more of a chance if they kept you up to date of the situation on a blow by blow basis but had a bit more downtime say 99% compared to a host who gives little information other than "we are down, looking at it" but has a 99.5% uptime record?

    Am just curious
    That's about a 7 minute difference in a day. If I had a website selling sunscreen and its 3am in december, I'd rather skip the blow by blow and get my beauty rest. Down 14 minutes instead of 7 wouldn't phase me at that time

  37. #37
    Join Date
    Feb 2006
    Location
    Kusadasi, Turkey
    Posts
    3,270
    Quote Originally Posted by andrewjab View Post
    Somehow (despite the servers being monitored by Nagios every few seconds) our customers still manage to somtimes submit a ticket when a web server (apache) has died before nagios picks it up. I often wonder how they notice soo incredibly quickly!
    We used to call it "client alert system". It's the most effective monitoring system. Even if the server goes down for 2 minutes at 6 AM in the morning, we always have a ticket letting us know

    Sometimes we considered removing the redundant monitoring software, since we already have the "client alert system".

    We suspect some of our clients uploaded themselves to their websites via FTP, so they always know when the server is down.
    Fraud Record - Stop Fraud Clients, Report Abusive Customers.
    █ Combine your efforts to fight misbehaving clients.

    HarzemDesign - Highest quality, well designed and carefully coded hosting designs. Not cheap though.
    █ Large and awesome portfolio, just visit and see!

  38. #38
    We suspect some of our clients uploaded themselves to their websites via FTP, so they always know when the server is down.
    I knew there was a secret to their fiendish way of always being one step ahead of us

    Its just so strange, I started to question whether 1 or 2 of our customers were sitting there refreshing with their other hand poised on the submit ticket option . . .


  39. #39
    Join Date
    Nov 2002
    Location
    USA
    Posts
    211

    web hosting time outs good or bad?

    Quote Originally Posted by Collabora View Post
    Interesting, no one has considered the time-frame involved.

    For matters under their control, a shared hosting environment should easily provide over 95% each month (that's 36 hrs/month downtime). If they can't do that you've got some incompetents working for you
    I have had customers with less the 15 min. in a month scream. It is hard to please everyone all of the time, no matter what you do everyone wants 100.9% uptime If anyone knows how please let me know.
    Cpwebhosting.net where the customer is first
    Great Affiliate Program Call 321-205-9003
    Earn on CPA up to $100.00 on Shared Hosting

  40. #40
    Join Date
    Mar 2009
    Location
    Santa Monica, CA
    Posts
    3,372
    Quote Originally Posted by rcrrich View Post
    I have had customers with less the 15 min. in a month scream. It is hard to please everyone all of the time, no matter what you do everyone wants 100.9% uptime If anyone knows how please let me know.
    If the 15 minutes was a random 30 sec per day in a month I doubt you would hear much about it. Also, keep in mind that what most of the contributors here call "uptime" is determined by a ping report, not site response. See my earlier post on the matter.

Page 1 of 2 12 LastLast

Similar Threads

  1. Replies: 45
    Last Post: 05-15-2007, 02:53 AM
  2. What Are The Causes Of Downtime
    By Ken7 in forum Web Hosting
    Replies: 12
    Last Post: 02-14-2006, 03:41 AM
  3. Customer complains about downtime...but no downtime according to log files
    By thomas.smith in forum Providers and Network Outages and Updates
    Replies: 6
    Last Post: 11-09-2004, 12:57 AM
  4. Downtime in DC2?
    By Urban Hitz Radio-NY in forum Dedicated Server
    Replies: 2
    Last Post: 09-29-2003, 02:01 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •