Dependencies   Leave a comment

What becomes a headache in a live system? Probably you might have guessed from the title, its dependencies – both internal and external. Specific internal dependencies are easy to monitor and maintain, when it’s an external, it’s a chaos.

What should one learn to deal with this? Mainly, one should evaluate the dependency. In other words we MUST test it and make sure that the communications and statuses are well formed and will not damage your system by changing them “on the go”. There are 100s of such dependencies ruining the beauty of the internet. #1 of them is the payment integrations with many online banking systems. These portals change their responses and features almost daily to comply with the financial industry standards and other security regulations and guidelines. Who will keep on monitoring them — well you should be, otherwise your precious customers will part from your web systems and online services.

Now you have evaluated the ‘standard’ of the external service, you should now be looking at our system to define ‘what will happen if this fails one day!’ – may be you are crossing your finger to never happened this, but there are times when the giants fail. So be prepared for that unfortunate day. Now, the first thing your developer should be doing is, check for the failed dependencies, and try to ‘catch’ them without going into a crash. Next, the important thing, the first glimpse of the dependency failure should be notified to the System Administrator. Also ask them to email/SMS the error code and the details of which dependency failed at what time with what error. Now the same thing should be logged with the whole dump of the session to analyze the failure.

Once your system is equipped with thee monitoring and alert modules, you should designate a team to deal with these situations. The team should regularly verify the status of the dependencies, get information about any policy or feature change related to them, get any updates related to its service and engage in an escalation process with these service providers.

Recently, there emerged many SaaS services that are free and are very useful. But, the free is always limited in some way, if you take Google Mail; there are 500 outgoing mails per day per box limit. Similarly there are limits line, CPU timing, request per day, request per host, maximum data length, and on. Your system should also watch for these limits and deal and alert accordingly.

The bottom line I one should depend on some reliable, standard services which as some support available, if it’s a free one track the limits.

Happy ‘In dependence’ Day


Posted October 9, 2011 by jsusmer in Uncategorized

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: