Howdy, Stranger!

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


Provider Communication - Feedback wanted
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.

Provider Communication - Feedback wanted

KuJoeKuJoe Member, Host Rep
edited November 2012 in Providers

First off, this thread is not about any specific provider. Do not turn this thread into one about this provider or that provider. This is just a thread where me, a provider is interested in what you, the client expects. I hope this thread will help other providers also but for the sake of this discussion please don't bash other providers. Thanks!

For example, if your VPS is offline for any reason how to you expect to be informed? Do you want an e-mail right away? Do you want an e-mail when it's back online? Do you want an e-mail if it's expected to be down for X more hours? Would an announcement be enough? How about Twitter updates?

I ask because I personally avoid trying to send mass e-mails except for the most important announcements but always keep our announcements updates (which mirrors to our websites, Twitter and our info site via RSS). The main reason for this is because we were previously using Google Apps for our e-mail up until a few days ago and they have a daily limit of 2000 e-mails with some crazy throttling after about 500 e-mails in a short span of a few minutes. We recently switched to SendGrid but I'm still hesitant to go the mass e-mail route.

So here's my question to all of you. What kind of communication do you expect from your provider and when?

Additionally, how many of you check announcements or utilize RSS feeds? I know a handful of people have thanked providers on here for keeping the RSS feed updated so I know some of you do, but I get the feeling it’s a small percentage.

Personally I always check announcements for my providers (Hostigation, BuyVM, and RamNode) but don’t utilize the RSS feeds.

Comments

  • MaouniqueMaounique Host Rep, Veteran

    I think you are asking the wrong crowd here.
    The regular LET ppl have tons of VPSes and an email would be best unless they already have another monitoring system like pingdom/nodeping, you get the idea, while the most addicted have their own grid keeping tabs on itself.
    I gave the VPSes to friends and they call me when there is a problem,then I go at provider to look for announcements and if there is none I ping them.
    There are few for myself but they are for development, testing, benchmarking so nothing vital there.
    Uncle posted on Twitter, now I also put an announcement in the forum if there is something unexpected and Uncle sends mails to ppl that are to be moved, "maintained" or something.
    So far nobody complained about these methods, but I am not sure everyone is happy, so, yeah, this thread will be interesting.

  • twitter with text alerts set up works best for me

  • klikliklikli Member
    edited November 2012

    First thing appears in my mind is to notify your customers in a timely manner. I often see providers announcing their maintenance on their website, but rarely through email. In fact, in my opinion all announcements posted at https://my.securedragon.net/announcements.php should be sent via email.

    Second suggestion from me might be some text alerts when there's some outage that is caused by your side, as such customers will feel they're informed. I do understand this is going to cost you but after all, such kind of announcements are only made when you make a mistake.

  • MaouniqueMaounique Host Rep, Veteran

    @klikli said: such kind of announcements are only made when you make a mistake.

    Actually, a mobo can burn even without a mistake, or a raid controller, god forbid the core switch... Some things cost 30 k a piece and it is not a mistake not to keep a spare.
    Plus, some numbers are fake, some networks are not connected, some ppl dont want to be waken up at 3 AM, so on and so forth.
    An email is the most I would accept to be bothered with, if that is so important you will redirect your email to some text or set your phone to receive them and notify you by a beep or something.

  • KuJoeKuJoe Member, Host Rep

    @klikli Now that we have SendGrid I plan to utilize e-mail more often than before, but text alerts does seem like a complicated solution to manage.

  • if your VPS is offline for any reason how to you expect to be informed? Do you want an e-mail right away? Do you want an e-mail when it's back online? Do you want an e-mail if it's expected to be down for X more hours? Would an announcement be enough? How about Twitter updates?

    my answer is yes to all above.

  • I think you should post an announcement inside WHMCS. That's where I look first if a VPS goes offline.

  • To me, the exact method doesn't matter too much. I'd rather that you're absolutely consistent though. Don't post some stuff to twitter, some via email, some on your announcements page - pick one method and stick to it. Or better yet, set up some kind of system that posts the same announcements to all of those with one click.

  • @klikli said: Second suggestion from me might be some text alerts when there's some outage

    As long as it's opt in only. I don't want to pay to recieve text messages because something messed up.

  • Nick_ANick_A Member, Top Host, Host Rep

    @NickM said: To me, the exact method doesn't matter too much. I'd rather that you're absolutely consistent though. Don't post some stuff to twitter, some via email, some on your announcements page - pick one method and stick to it.

    Agreed--always use the same base method (email or Twitter or something) and then contact through other means as needed.

  • I would expect an email for planned maintenance, outages, interruptions, transparency, updates, etc and have that information mirrored on the provider site.

    If the VPS is down and I don't have an email from them, I go to the providers site to file a ticket, an announcement there can sometimes circumvent my ticket.

    I could care less about RSS, Twitter.

  • This might not be par for the course for an LET member, but since I run services on my VPS that are for others and not just myself, I am beholden to those others to make sure things work. Because of that, I like as much communication and openness as possible. I always like to know what's going on. I recently switched providers (in fact, I'm now using both your services, @KuJoe, and BuyVM's) and my choice was largely based on my past experience communicating with said providers. The speed at which SecureDragon responded to my sales inquiries (on a weekend no less) and the knowledge that I can often find @Fransisco or @Aldryic in IRC gives me definite comfort.

    Also, I completely agree with @NickM, keep it consistent. Send out the same info via e-mail, Facebook, Twitter - in fact, why not do all of them? Why not let users opt-in to alert e-mails if they want them, but trust the users who don't opt-in to find the info they need elsewhere, be it via WHCMS (which should always be up-to-date; not enough providers utilize its ability to record issues), or some social media platform?

    As far as what communication to have, I absolutely want to know not just if there's a possibility my VPS will go offline, but if it is offline and I might not have noticed. I would especially like to know if it's predicted to be offline for some time more, and I'd generally like to know why - as much as you're able to tell me. Particularly when there are serious issues, transparency is absolutely key. Recent incidents with various providers, ranging from very sophisticated and well-setup ones to kiddie hosts, have shown that when communication is inconsistent or conflicting, people assume the worst. In the event that you must be inconsistent (something you thought was working actually isn't, or vice versa), make a note of it and apologize - then it doesn't feel as though it's being ignored or covered up.

    I myself don't use RSS feeds, but I check WHCMS/Twitter/Facebook for the providers I use. I also lurk here a lot, and chat in the various IRC channels associated with the site to whatever degree.

  • Post to your website is sufficient for me.

  • @Ivraatiems said: let users opt-in to alert e-mails if they want them, but trust the users who don't opt-in to find the info they need elsewhere, be it via WHCMS (which should always be up-to-date; not enough providers utilize its ability to record issues), or some social media platform?

    Pretty much this.

  • jarjar Patron Provider, Top Host, Veteran

    I can tell you what I do, because it fits what I would want as a customer. The second I notice an issue that can't be resolved within 20 minutes, I e-mail everyone with an active service on the node. I try to provide 24 hour updates if the issue persists for days. Of course, no issue should last for days, but I have a node in FDC Denver. Long story, I'm sure more will be said about that elsewhere in a few days, but that's how I approach it. I've not had a complaint, though obviously I'm not having to send out e-mails every week over it.

  • This comes down to establishing a policy, an emergency policy as a provider. Meaning, it should be documented and ideally automated. But should also contain manual steps to do the notifications in case of larger failure/outage.

    1. Email. Start with that. You already do billing and other communications via that.

    2. Post to social media (whatever the ever changing fashion there is --- Twitter, Facebook, etc.).

    3. Create a section on your homepage for emergency news and have the info there. You can later pull that off/down when no longer active.

    4. Put notifications in your control panels and or billing.

    Since this is your business, I expect that you have a more watchful eye on your servers and bandwidth than I do.

    If you take these steps and automate the process can be real painless. It should reduce the barrage of tickets that fly in and later waste time to process.

  • craigbcraigb Member
    edited November 2012

    Hey @KuJoe - excellent question, good to see you asking.

    I strongly prefer email but really don't like emails that either assume I know what node I'm on (too many LEBs, so if I know the node, it's normally a bad sign - too many outages) or aren't relevant to my service/node. So some way to only notify customers on an affected node is important to me. I like as mich notice as you can give for scheduled outages (shouldnt be less than 24 hours unless circumstances dictate otherwise). Also: if a provider gets hacked, I want to know within 4 hours of you knowing (sooner preferably) and then twice daily email updates if possible until things return to a steady state. Twitter and RSS mirroring the headlines with a link for the content is definite nice to have. I only connect with friends with Facebook.

  • KuJoeKuJoe Member, Host Rep

    Very helpful information. I've never used SolusVM's built in mass mailing feature but I may start to utilize that for specific nodes (we used to do this in WHMCS but recently made a change for better node management). I am also working on setting up a cluster for WHMCS (similar to our current 8 server cluster but hopefully beefier) since I've been having a lot of fun setting up replicated MySQL servers the past 48 hours. :D

    Keep the ideas coming, I'm hoping other providers are benefiting from this also. :)

  • I would be hesitant as a hosting company on Facebook. Go check out Virpus. Better hire a guy just to do Facebook all day and make sure he doesn't get up to use the bathroom because during some downtime outages, you will get flooded on there with angry comments

  • Well @bamn, agreed about Facebook, but rather deal with folks there instead of having 500 messages in WHCMS or whatever you use.

  • LeeLee Veteran

    Personally all I want is an email telling me something is wrong in the first instance, I get my emails wherever I am so it's a good first point of contact.

    Within that first email I then want to know where I will get my updates, twiiter, facebook or status page on your site. Just tell me the one place you will put updates so when I want to know what is happening I can go there as I need to and if I have questions I can raise a ticket or send an email.

  • I've been thinking about this recently, as I have a lot of VPS's, at a lot of providers.
    I'm putting together my own automated grid, to manage my services on the VPS’s. But it would be useful if there were some standard code reference, something like: Ten-codes (example: http://spiffy.ci.uiuc.edu/~kline/Stuff/ten-codes.html). That way when an email or twitter message is sent out, I could have my system implement policies based on the new intel. Move stuff around, make emergency backups, move services over to different nodes etc.

  • I agree with your idea @bit.

  • Just off the top of my head, if there is more interest we can start a separate thread.

    Place holders:
    [provider]
    [network] or [location] or [DC]
    [node]

    [date] [from date] [to date] - Date in format: yyyymmdd
    [start of window] [end of window] - time, in a standard time format: e.g. UTC hhmm
    [old end of window] [new end of window] - time, standard time format: e.g. UTC hhmm

    [all clear] - all clear signal

    [normal but be warned] - back to somewhat normal, but there could be problems, not a full all clear signal.

    Examples - (needs ten-codes for the codes):
    wobbly/unstable node - [node]
    under DDOS attack - [network] or [location] or [node]
    network problem - [network] or [location] [node]

    planed maintenance - [node] [date] [start of window] [end of window]
    emergency maintenance - [node] [date] [start of window] [end of window]
    maintenance window extended - [date] [old end of window] [new end of window]

    warning potential problem - [from date] [to date] or [normal but be warned] or [all clear]
    prepare for shutdown - [date] [start of window] [normal but be warned] or [all clear]

    ....

    The last ones from "warning potential problem", could be useful in situations like Hurricane Sandy.

    prepare for shutdown - a light/power out situation e.g. generators run out of fuel by 2:30 am, wait for signal

  • I'd prefer if every provider ran a simple, off-network page with RSS. Then I have one place to go (my RSS reader) to see maintenance / outage announcements for every provider I use. Make a second feed for news/offers/adverts and I can choose to subscribe or not.

    Email is overused. Spam filtering is becoming stricter all the time and many providers fail to correctly configure their mail systems, or shift things around and forget to update the mail config.

    The client has a responsibilty (IMO) to seek information. The provider is responsible for publishing it in an easy-to-get format.

  • SecureDragon customer here:

    For me, I think the best method of being alerted of anything is mostly email. I'm not for sure if you do it, but I don't recall getting a scheduled maintenance email in the past. I was using the server, and it went offline. Of course, your website had the maintenance schedule posted so I then knew not to worry and knew not to bother you guys with a ticket, but an email beforehand would be nice.

    And honestly, TXT message to phones wouldn't be a bad idea either. So long as it wasn't a txt for every little update or notice. For example, you could email your clients 48 hours before planned maintenance, txt them 24 hours before (or less, like 12, 8, or 4). I'd only want txted if it was an emergency, downtime, or if something needed to be seen by me ASAP.

  • Email on scheduled downtime. Now and then unexpected downtime, I don't care as long as it's not too frequent or long since damage already done.

Sign In or Register to comment.