MightWeb is a small web hosting company that was founded by Marcus Frolander in January of 2015. I’ve been on a Starter package since May of 2016. In August of 2016 MightWeb merged with ImbaHost and went from being a one-man operation to a team of at least five. Some things that have really stood out for me with MightWeb are their transparency and high standards for quality of service.
Since I've been using MightWeb, my hosting account has been through two server migrations. I was originally on the s103-carrier server and then was moved to the s104-carrier server, both in Carrier-1 facility in Dallas, TX (migration was completed on 1-27-2017 with zero downtime). I'm currently on s101-cdc603 server in their new location at the NCDC603 facility in Chicago, Illinois (server migration completed on 06-25-2017 with no issues).
On top of always informing customers of any planned maintenance well in advance, MightWeb also sends out "Reason for Outage" emails whenever there has been any unplanned downtime/issues. Sometimes the details are outlined in an PDF attachment, here's an example from 2-20-17. The client area shows the current status of servers, any past incidents that have been resolved, as well as any future scheduled maintenance (https://mightweb.net/status/). They also send out newsletters usually on a monthly basis with reflections/insights, recent improvements that have been made, and highlights of any upcoming changes.
Speed Testing - Fast Load Time
As mentioned above, I'm currently on a server located in Chicago, IL. Here's a speed test from the Pingdom Dallas, TX location. Less than a second (963 milliseconds) load time for a 2.1 MB homepage is fast.
Dallas Pingdom Speed Test — Load time: 963 milliseconds
MightWeb provides shared hosting, reseller hosting, VPS (Linux and Windows), and dedicated servers. The shared hosting "Starter package" I'm on allows for hosting 2 websites and 1 MySQL database. I found it odd at signup that this plan offers an add-on domain but no option for an additional MySQL database. I debated between the Starter and Advanced packages but ended up opting for the Starter plan despite the useless add-on domain since I don't have any static sites and just considered it a single domain plan.
There were some long instances of downtime last year that were mostly due to planned maintenance. MightWeb ended up issuing a refund for a full year of service in September because of not meeting their uptime standards. More recently their uptime has been significantly better.
June 2016 - 99.95% Uptime (20 minutes of downtime) Psychz NOC confirms a datacenter-wide network outage on June 17.
July 2016 - 99.48% Uptime (234 minutes of downtime) Network Hardware Maintenance -- code & configuration updates/changes to cabinet switches on July 6.
August 2016 - 99.96% Uptime (17 minutes of downtime)
September 2016 - 99.56% Uptime (190 minutes of downtime) 3 hours down on the 12th and full year's refund was provided (more details on this down below).
October 2016 - 99.99% Uptime (6 minutes of downtime)
November 2016 - 99.52% Uptime (206 minutes of downtime) Server maintenance to increase disk capacity and I/O speeds of the server on November 10th.
December 2016 - 100% Uptime (0 minutes of downtime)
January 2017 - 100% Uptime (0 minutes of downtime)
February 2017 - 99.57% Uptime (172 minutes of downtime) The upstream provider null-routed the main shared IP address of the server on February 20th. Null-routing should have not occurred in the first place.
March 2017 - 100% Uptime (0 minutes of downtime)
April 2017 - 100% Uptime (0 minutes of downtime)
May 2017 - 99.91% Uptime (41 minutes of downtime)
June 2017 - 100% Uptime (3 minutes of downtime)
July 2017 - 100% Uptime (0 minutes of downtime)
August 2017 - 100% Uptime (0 minutes of downtime)
September 2017 - 100% Uptime (0 minutes of downtime)
October 2017 - 99.72% Uptime (127 minutes of downtime) Network experienced re-occurring packet loss.
November 2017 - 99.95% Uptime (23 minutes of downtime)
December 2017 - 99.96% Uptime (16 minutes of downtime)
January 2018 - 100% Uptime (0 minutes of downtime)
On September 12th, 2016 after 3 hours of downtime that day the owner Marcus issued a full refund for a year of service.
Marcus: We are getting in touch with you as we've unfortunately noticed that we've failed to maintain our uptime guarantee for the past three months. We strive to maintain a high level of service, but a chain of unfortunate events have resulted in our network not being as stable as it normally is.
As a result of this, and as a result of your uptime not being up to par with what we expect from ourselves, we are hereby issuing a full refund of your services. Your services will remain active of course - so see this as our way of apologizing for the recent issues. Rest assured that we are working diligently to resolve them.
If you have any questions or concerns, please let me know!
I felt that a full year's refund wasn't called for...
Me: I really appreciate the refund, thank you. I would like to go over the statement "we've failed to maintain our uptime guarantee for the past three months".
From my records I had less than 20 minutes of downtime in August.
The downtime in July (around 4 hours) was on July 6th which I believe was related to planned maintenance on a switch.
September (around 3 hours).
At the most two months of hosting account credits for July ($2.59) and Sept ($2.59) would have been sufficient enough for your uptime guarantee. Thank you for going well above and beyond any SLA but if you'd like to do the hosting account credits instead please let me know.
I hope the network issues are resolved moving forward and thanks again.
Marcus: Our monitoring (SrvTrack) measured a server-wide uptime of 99.86% during August as a total. Granted we have it set fairly strict where it generally accounts for downtime on the second consecutive connection failure - but I'd still dare to say the uptime was less than our guaranteed 99.9%.
As for the other two occasions - whilst one was planned, the total downtime for these two events are in my book enough to warrant a proper refund. That is especially considering that they were in such a short proximity to one another. As I mentioned - we set high standards for what we should provide both in terms of support and uptime. When we fail either, it only makes sense that you are properly credited for it.
There is no need to deposit anything in terms of account credit - simply consider this my (and our) sincere apology for failing to hold up to our guarantees. I assure you we're working out solutions in co-ordination with the data center to get back on track, and prevent these issues.
In the year and some change that I've been a customer of MightWeb it is clear to me that they won't settle for anything but perfection. If things don't go as intended they own up and provide open and transparent communication with their customers with well thought out procedural actions to fix any issues with prevention in mind. The times where things have gone awry they have disclosed in full detail what caused any failure whether on their end or by providers they have entrusted within their infrastructure. The founder of MightWeb, Marcus, truly cares about the satisfaction of customers. If you are looking for an honest web hosting provider with top notch performance and proficient and friendly support MightWeb is definitely worth your consideration.