I have a website: Lee-Lo. Small and personal. Very personal. Since I was.. let me remember.. 11 or 12 years old I liked collecting sayings and quotes. From famous people, from unknown people, all kinds. I used to write them in a notebook by that time. What’s really nice about sayings and quotes is they educate you, at least most of them. And I loved reading them, from time to time. I think they helped me become a good person. Anyway, the time passed, some moment in my life I gave up collecting them, until about one year ago when I decided to make this website, to collect memes and photos with sayings, most of them from facebook which is full of such, and build sort of a personal collection, online. I love them now as much as I loved them when I was 11 or 12.
It’s not much of a website. It doesn’t have traffic, it doesn’t rank high on Google or on other search engines, it’s nothing more than a personal collection online. Consequently, decided it’s not the case investing lot of money in web hosting. So I made my choice on Planet Hippo. They promote themselves as:
– Low Cost Web Hosting – High quality Website Hosting at an affordable price – you won’t find better elsewhere.
– High Speed Network – One of the fastest, most reliable networks in the industry with 99.99% uptime guarantee.
– 24/7 Support – Guaranteed response within 30 minutes and total resolution within 4 hours – around the clock.
And price, huh, £1/month, excellent. Best choice, isn’t it? At least at first sight. I was not ashame I choosed them. Since they promote themselves as low cost web hosting, 99.99% uptime guarantee, 24/7 support, they have to assume their promises, right? You can’t lie about yourself. Otherwise people will discover it, and will start talking about it.
Getting to the point, never had a problem with short service interruptions: updates, upgrades, resets, they are all normal. So 1 minute break, 2 minutes break, let’s say even 5 minutes interruption I can understand, when they happen once in a while. But know what? I have an account with another service I fully recommend, Uptime Robot, a service that monitors websites at 5 minutes intervals. If it finds the website is safe and sound, it checks again in five minutes. In case it discovers the website is down, it will check every minute until it’s back online. So, every time the website goes down, I receive an email containing a notification on the situation. Every time the website gets back online, I receive another email containing the good news. Of course, they may be plenty of times it doesn’t notice a problem with my website simply because they’re too short and the monitor doesn’t have the chance to detect them. It’s blind, so am I.
So, if service interruptions are, once a month, and short, I don’t have a problem. The interesting part comes when it detects, or start detecting multiple interruptions. And longer than expected. Or considered acceptable.
For a long period of time I hadn’t monitored (I started the website and with Planet Hippo on September 27, 2015), hadn’t any hosting problems, or didn’t care about interruptions. I don’t remember when exactly it happened but some time during 2016 I started noticing the website goes down more often. When I felt it wasn’t going in the right direction I started opening support tickets. Like on October 2, 2016 when I opened a ticket claiming the website is down, didn’t know for how much time already. 40 minutes later I received the confirmation the website was live again. No explanation on what the problem was but anyway wasn’t interested in finding out this detail.
On November 17, 2016 I told them in another support ticket the website was down for more than 1 hour, kindly asking them to check and fix it. It went back live after 1 hour, 30 minutes and 39 seconds (thanks Uptime Robot for the info), so I asked the Planet Hippo support team: “Now, please tell me what was the reason the websites went down and how did you solve it“. Curious about their answer? “It was a temporary network issue and resolved it now“. So I replied: “Please, what kind of temporary issue? You can imagine, I want to know, and I guess I can understand it, even if you explain it technically“. The answer: “It was a temporary network failure issue. The network is not reachable at that moment“. Since I was involved in a chat ealier, replied: “Network failure? Do you know what your colleague in Live chat told me? ‘This is due to delay in propagation. Please wait for a while for the propagation to be complete. This is normal for a new domain.’” But what kind of new domain? 😉 I was with them for more than 1 year already.
That answer made me angry so, starting that day, never deleted the emails received from Uptime Robot anymore. I knew that if they continue, I might build a case with that information.
And it was the case, the next day, November 18, 2016. So, new support ticket: “After yesterday interruption of 90 minutes, my websites are down again, repreatedly:
from 19:47 to 20:28 ~ 42 minutes down
from 20:33 to 20:37 ~ 4 minutes down
from 20:42 and still down.
What happens with you guys? Looking forward for your fix.”
More than 3 hours later, I got their final answer: “The issue has been fixed and both your websites are working well. Let us know if you need further assistance.” Now you know I had another website hosted at Planet Hippo (I moved it to another hosting company in the meantime).
The 3rd ticket opened on this matter was on November 30, 2016: “My website is not working, is down. Why? And why so often?” Their answer, 10 minutes later: “After checking, your website is working fine.”
Stopped asking for support via support tickets but here’s the history, Planet Hippo‘s glorious history of downtimes. Now remember, all those downtimes might be longer than reported by Uptime Robot, because it checks the website every 5 minutes while it can go down earlier, then approximately every minute until it’s back live.
2016, November 17: 21:06-22:37 = 1 hour, 30 minutes and 39 seconds
2016, November 18: 19:47-20:28 = 41 minutes and 30 seconds
2016, November 18: 20:33-20:37 = 3 minutes and 29 seconds
2016, November 18: 20:42-23:15 = 2 hours, 32 minutes and 58 seconds
2016, November 18: 23:31-00:16 = 45 minutes and 29 seconds
2016, November 19: 01:58-02:01 = 2 minutes and 51 seconds
2016, November 19: 04:46-04:47 = 1 minutes and 1 second
2016, November 19: 07:47-07:50 = 2 minutes and 22 seconds
2016, November 19: 13:45-13:47 = 2 minutes and 22 seconds
2016, November 20: 04:53-06:19 = 1 hour, 26 minutes and 21 seconds
2016, November 23: 04:40-04:41 = 1 minutes and 1 second
The same day I received a notification from Google Search Console: Googlebot found an increase in server errors on https://lee-lo.co.uk/
Not nice. And the story continues.
2016, December 02: 16:38-16:39 = 0 minutes and 58 seconds
2016, December 20 13:29 – December 21 06:02 = 16 hours, 32 minutes and 55 seconds
2016, December 24: 04:41-04:42 = 1 minutes and 2 seconds
2016, December 27: 08:39-08:47 = 8 minutes and 32 seconds
2016, December 27: 16:38-16:39 = 1 minutes and 0 seconds
2016, December 29: 04:45-04:46 = 1 minutes and 0 seconds
2016, December 30: 12:42-12:43 = 1 minutes and 1 seconds
I wonder how many other short interruptions (under 5 minutes) went undetected by Uptime Monitor. They have a premium plan checking the website every single minute, but I didn’t want it, thought it’s not worth it, I didn’t want to get even more upset. Happy new year 2017? You bet, no.
2017, January 02: 12:41-12:42 = 1 minutes and 1 second
2017, January 03: 12:47-13:17 = 30 minutes and 20 seconds
2017, January 04: 04:40-04:41 = 1 minutes and 0 seconds
2017, January 05: 20:38-20:39 = 1 minutes and 0 seconds
2017, January 06: 04:45-04:47 = 2 minutes and 22 seconds
2017, January 07: 04:42-04:43 = 1 minutes and 1 second
2017, January 12: 12:43-12:44 = 1 minutes and 0 seconds
2017, January 12: 20:40-20:41 = 1 minutes and 1 second
2017, January 13: 08:42-08:43 = 1 minutes and 1 second
Do you notice there are so many short interruptions around 04:40, 12:40, 12:40, 20:40? I suppose it’s something the guys at Planet Hippo are quite aware of it, maybe normal, maybe not, I don’t know, I guess there are some processes scheduled every single day but with devastating effect of putting websites down. Well, as long as they’re short, I can eat it. But 30 minutes, 1 hour, 16 hours(!), that’s outrageous.
2017, January 16: 04:53-04:56 = 2 minutes and 21 seconds
2017, January 16: 12:42-12:43 = 1 minutes and 2 seconds
2017, January 16: 20:40-20:41 = 1 minutes and 1 second
2017, January 17: 16:40-16:41 = 1 minutes and 1 second
2017, January 20: 12:43-12:44 = 1 minutes and 1 second
2017, January 21: 12:42-12:43 = 1 minutes and 2 seconds
2017, January 22: 16:41-16:42 = 1 minutes and 1 second
2017, January 24: 16:42-16:43 = 1 minutes and 1 second
2017, January 24: 20:43-20:45 = 1 minutes and 23 seconds
2017, January 25: 04:51-04:52 = 1 minutes and 18 seconds
2017, January 25: 20:42-20:43 = 1 minutes and 0 seconds
2017, January 26: 00:41-00:42 = 1 minutes and 1 second
2017, January 26: 04:52-04:53 = 1 minutes and 0 seconds
2017, January 26: 08:43-08:44 = 1 minutes and 0 seconds
2017, January 26: 11:17-11:38 = 21 minutes and 8 seconds
2017, January 26: 16:42-16:43 = 1 minutes and 2 seconds
2017, January 26: 20:43-20:44 = 1 minutes and 2 seconds
2017, January 27: 04:56-05:00 = 4 minutes and 54 seconds
2017, January 27: 12:48-12:49 = 1 minutes and 22 seconds
2017, January 27: 16:11-16:14 = 2 minutes and 45 seconds
2017, January 28: 00:44-00:45 = 1 minutes and 0 seconds
2017, January 28: 04:55-04:56 = 1 minutes and 1 seconds
2017, January 28: 16:47-16:48 = 1 minutes and 2 seconds
2017, January 28: 23:30-23:32 = 2 minutes and 24 seconds
2017, January 28: 23:37-23:44 = 6 minutes and 38 seconds
2017, January 29: 04:50-04:53 = 2 minutes and 22 seconds
2017, January 29: 12:48-12:49 = 1 minutes and 0 seconds
2017, January 30: 08:43-08:46 = 2 minutes and 22 seconds
2017, January 30: 12:46-12:48 = 2 minutes and 21 seconds
2017, January 30: 16:46-16:47 = 1 minutes and 0 seconds
2017, January 30: 18:09-18:39 = 30 minutes and 1 second
2017, January 30: 20:45-20:46 = 1 minutes and 24 seconds
2017, January 31: 00:41-00:43 = 1 minutes and 22 seconds
2017, January 31: 08:44-08:57 = 13 minutes and 4 seconds
2017, January 31: 12:47-12:48 = 1 minutes and 0 second
2017, February 01: 08:43-08:46 = 3 minutes and 22 seconds
2017, February 01: 20:42-20:44 = 1 minutes and 22 seconds
2017, February 02: 05:01-05:02 = 1 minutes and 8 seconds
2017, February 02: 16:47-16:48 = 1 minutes and 1 second
2017, February 02: 20:44-20:45 = 1 minutes and 1 second
2017, February 03: 04:57-04:58 = 1 minutes and 2 seconds
2017, February 03: 08:44-08:46 = 2 minutes and 22 seconds
2017, February 03: 12:46-12:47 = 1 minutes and 2 seconds
2017, February 03: 15:58-15:59 = 1 minutes and 2 seconds
2017, February 03: 16:44-16:45 = 1 minutes and 21 seconds
2017, February 03: 17:36-17:38 = 2 minutes and 22 seconds
2017, February 03: 19:15-
At this point I couldn’t stand an started a chat with them:
Me: My website lee-lo.co.uk went down 11 times today, and the 12th started more than 1 hour ago, at 19:15. When will you stop with all these problems?
HP: Let me check. The issue was due to sudden load spikes on the server
Me: I don’t care
HP: We are currently working on this issue
Me: I don’t care either
HP: This will be fixed soon
Me: I want my website be live as you promise 99.9% uptime. And it takes you more than 1 hour? Come on, I bet you had no idea about it until now
HP: We are aware of the issue. We are working in this issue
Me: Almost 2 hours downtime is not an issue, it is a problem. A problem repeating over and over again.
HP: This issue will be fixed soon. Please be patient
That was the moment I decided to write this post, to share my experience with Planet Hippo and to ask you am I freaking out? It is normal and should I stay quite?
The story continues anyway:
The next morning when I woke up, the website was still down. New chat:
Me: My website lee-lo.co.uk is down since last evening 19:15
HP: Let me check. Our admin is working on the issue
Me: Last evening you told me here on chat “The issue was due to sudden load spikes on the server. We are currently working on this issue. This will be fixed soon”. If you have a problem you can’t solve, at least offer alternative solutions. Your website is live. Mine isn’t. Why?
HP: We are currently working on this and we hope this will be fixed soon.
Me: Ok, I ask you again: Your website is live. Mine isn’t. Why?
HP: We have different server’s and our website is not hosted on the same as your
Me: Ok, we have a solution, just put it into practice: move my website on the same machine that hosts your website. Otherwise I will think your 99.99% uptime guarantee is for your website only. Will you do it? Will you move my website?
HP: We’re really sorry to inform you that we also don’t have access to the server.
Me: I don’t understand, you don’t have access to your own server, right?
HP: The server is currently down and our senior engineers are working on it to bring back the server as fast as possible.
Gave up. But as today was the first time to see all those short downtimes alltogether, a question raised into my mind. Leading to a support ticket:
Hi guys, as I already told you in previous occasions, I was monitoring the status of my website with Uptime Robot (you can see below the list of downtimes) and you never complained about it, so it’s fine.
I have one question related to the fact these downtimes are most of them, linked to some exact moments during the day: 04:40, 08:40, 12:40, 16:40, 20:40, 00:40. So the question is: what kind of processes are you running every single day? They are putting my website down every single day, more times per day, and I suppose they are more but Uptime Robot can’t detect them all because they’re too short to fall under the 5 minutes ping on my website. I am also sure they affect all websites hosted on the same server with my website. If only my website affected – it should have been the same time, every day. While I can see sometimes it happens at 20:33, 16:38, 04:45, 04:53 to give some examples, meaning the process reach my website when it’s done with other websites.
So, as I want to understand this server behavior, what kind of processes are you running every single day, every 4 hours, leading to interruptions in the hosting service? Thank you.
They replied: Hi, We’re sorry for any inconveniences this have caused you. At the moment we have some issues with the server and our senior engineers are working on it to bring back the server. We’ll get back to you with an update as soon as possible.
I replied: Hi, I was not writing you about this downtime that started yesterday evening. I can understand it. And you.
My question is about that really long list of short downtimes. The current downtime only determined me to make the list and to realize something’s going on, every single day, every 4 hours, that gets my website (and I suppose other websites as well) down.
So, here’s the question again, please: what’s that all about these repeated short service interruptions? Please don’t avoid the question, give me a straight answer. I want to understand, not to judge.
Later on same day, the answer from Planet Hippo:
We apologize for the late reply. Regarding your question of repeated short service interruption, it occurred due to overload on the server. We are working on resource upgrade. We will get back to you with the updates once this issue is resolved. Mean while your patience is much appreciated.
Simply one can’t talk with these people who try to hide real answers. Overload on the server as they say is simply a lie. It can’t happen every 4 hours, precisely, every single day.
At this point I saved a copy of Planet Hippo’s Terms and Conditions, as of January 4, 2017. Just in case.
And we got news from Uptime Robot, so let’s resume:
2017, February 03 19:15 – February 04 12:33 = 17 hours, 17 minutes and 32 seconds
2017, February 04, 12:38-14:25 = 1 hour, 46 minutes and 39 seconds
2017, February 04, 14:30-17:51 = 3 hours, 20 minutes and 33 seconds
Finally! The website is live again, with a “nice” answer from Planet Hippo: “The issue has been resolved and your website is loading fine now. So please check the same from your end and let us know if you need any further assistance.”
2017, February 04, 18:06-18:08 = 2 minutes and 25 seconds
2017, February 04, 20:44-20:45 = 1 minutes and 21 seconds
2017, February 05, 05:00-05:04 = 3 minutes and 27 seconds
2017, February 05, 07:37-
A new 7+ hours downtime makes me chat with them, again (trust me, I feel no pleasure, but I think I deserve some answers):
Me: My website lee-lo.co.uk is down since 7:37 this morning. You previously said you solved the problem with your server. What’s now?
PH: We are currently working on this issue and it will resolve shortly.
Me: Shortly you say? In the last 48 hours my website was down 22 hours.
PH: Sorry for the inconvenience caused. We are solving this issue with top priority basis.
I realized I was wrong: in the last 48 hours the website was down more than 30 hours, missed to add the downtime today, 7 hours and still counting.
2017, February 05, 07:37-15:45 = 8 hours, 8 minutes and 6 seconds
2017, February 05, 15:50-16:00 = 10 minutes and 23 seconds
2017, February 05, 16:35-16:37 = 1 minutes and 22 seconds
2017, February 06, 04:41-04:42 = 1 minutes and 1 second
I don’t loose money if my website is down. Other websites hosted on the same server probably (surely) do, if they are affected by the same server behaviour (and, based on what the guys at Planet Hippo tell me, all the websites sharing the same server with me are down since the server looks like shut down).
I don’t loose rankings on Google or other search engines if my website is down. Other websites hosted on the same server probably (surely) loose.
The experiment (I now consider it an experiment) continues. I will definitely move the website to another hosting company, sooner or later, the money paid to Planet Hippo for hosting and for SSL certificate is not so much.
February 12, 2017: I got an automated email from the server: “The user has almost reached their disk quota. The account currently uses 82.51% (844.94 MB/1 GB) of its capacity. Remove some files from this account, or ask the system administrator to increase the account’s disk quota.”
Well, it’s as simply as removing some files from the server (I don’t want to upgrade yet) so let’s head to the cPanel. With Planet Hippo, you don’t have an username and a password to login to your website’s cPanel (eg. example.com/cpanel). Or at least I don’t know how to activate it. So I take the usual way: Client Area login, then select the hosting service and from there Login to cPanel. But surprise: Planet Hippo’s own website doesn’t work. Don’t know since when but I’m curious. Probably I will never find out. Anyway, F5 – nothing. CTRL+F5 – still nothing. The good thing is my website loads fine. This makes me think to avoid asking them to move my website on the same server that hosts Planet Hippo’s website. About 10 minutes later, still no Planet Hippo website. I will try again in a few hours.
2017, April 19, 04:57-06:00 = 1 hour, 2 minutes and 23 seconds
2017, April 21, 06:14-06:27 = 12 minutes and 24 seconds
2017, April 25, 04:47-05:17 = 30 minutes and 25 seconds
Feel free to share your opinion on this matter, it will be highly appreciated.