Most of the customers who have us do websites and web applications for them get us to host it all for them. Although we have to maintain the servers, this is usually a bit less faffy and more reliable than dealing with a third party. However, sometimes customers already have hosting, and want to carry on using this, which is fine too.
Today, I'm slightly confused by a relatively large third party hosting company who shall remain nameless. One of our customers reported that their web application had stopped working and was reporting database errors. Some investigation shows that their hosting provider has been doing upgrades (in the middle of the working day) and have shut down the PostgreSQL server... and it's been down for hours.
The control panel shows "This server is currently receiving security and feature updates. If you experience any problems, please try again later when this message has been removed. Thank you for your patience." - that's it, no ETA or anything. It also seems likely that they never informed their customers about this (but I can't actually guarantee that since I'm not their customer).
Is this really the expected behaviour of a reputable hosting company? I know we are very careful to minimise disruption and wouldn't dream of taking whole services offline for hours on end...
Wednesday 26 March 2014
Tuesday 25 March 2014
User guides
I've been spending some time writing user guides for the new Iceni 2 system (http://www.opendium.com/user_manuals/iceni). Yes, I know it might come as a shock that we're actually suppling documentation! :)
Its surprising how time consuming it is to write this stuff, put together screen shots, etc. even though this user guide isn't especially large: the aim is not to provide a 2000 page definitive guide that no one will ever read; instead we're doing reasonably short documentation that shouldn't take too long for someone to read through and gain a basic overview of how everything works. Eventually we'll also augment it with a knowledgebase of how-to documents explaining how to achieve things that people frequently ask for help with.
Although Iceni 2 has had a year's worth of development work put into it, many of the concepts date back to the original Iceni which has been running many customers' networks for years, so this documentation may even be of interest to those customers (who will, of course, eventually be migrated onto the shiny new system). Its surprising to think that we started developing the original Iceni back in 2005, and looking back on it we seem to have got most of the fundamentals about right right from the start.
While I've been doing the docs, we've also been rolling out our first Iceni 2 servers to customers. That seems to be largely going ok, although as always you do end up finding a few niggling bugs that never showed up during testing. So far they've been all quick to fix.
Its surprising how time consuming it is to write this stuff, put together screen shots, etc. even though this user guide isn't especially large: the aim is not to provide a 2000 page definitive guide that no one will ever read; instead we're doing reasonably short documentation that shouldn't take too long for someone to read through and gain a basic overview of how everything works. Eventually we'll also augment it with a knowledgebase of how-to documents explaining how to achieve things that people frequently ask for help with.
Although Iceni 2 has had a year's worth of development work put into it, many of the concepts date back to the original Iceni which has been running many customers' networks for years, so this documentation may even be of interest to those customers (who will, of course, eventually be migrated onto the shiny new system). Its surprising to think that we started developing the original Iceni back in 2005, and looking back on it we seem to have got most of the fundamentals about right right from the start.
While I've been doing the docs, we've also been rolling out our first Iceni 2 servers to customers. That seems to be largely going ok, although as always you do end up finding a few niggling bugs that never showed up during testing. So far they've been all quick to fix.
Subscribe to:
Posts (Atom)