You are browsing the archive for monitoring.

Photo Caching on Social Sites

6:34 am in infrastructure management, miscellaneous, monitoring by Matt Jenkins

I read the bbc news story today – which reports that a group at Cambridge University found that the direct URL to uploaded photos on various sites continued to exist 30 days after they deleted them.

Is this news?  These sites have several hosting challenges, one key area is caching.

Read the rest of this entry →

Disaster Recovery

11:24 am in monitoring by Matt Jenkins

The floods in northern England last week got me thinking about disaster recovery plans and scales of risk management etc (I’m and interesting guy really), and how my work would cope in a similar situation. Read the rest of this entry →

Site Monitoring

11:04 pm in monitoring by Matt Jenkins

I’ve been looking to proactive monitor the various websites with which I work.

To start with, I selected WebSitePulse.com to monitor the availability and load speed of the sites. The service is very simple, after a step by step configuration process the site will load a url (or ping, tracert, ssh, smtp & many others) and do a check for a predefined string on that page, you can select different locations for the check, and different check intervals. Over time, this will provide a good indication of uptime for that page.

I also setup a full web transaction test – where a transaction through your site can be pre-defined, and the monitor will run through the transaction recording page load speeds and link status. This is important because broken images, invalid http links and old failing pages hurt your SEO. It’s also nice to be able to see which pages take the longest to load. Is it those lard filled photos? That weighty database query? or the third party blogging tool you’re using (*cough*)?

The next stage is to understand when the problems are happening, and to get a record of the indicators for each server during that time. For this I use a combination of Serverscheck and Windows Perfmon. Serverscheck is a bit rough around the edges, but has a wealth of functionality for monitoring servers including SNMP, Windows Health Checks, Alerting via email, SMS and voice call. It produces simple graphs with the data it stores & can show the servers resource usage for any given period, which is also what I use perfmon for, although only during debugging periods (as it’s quite resource heavy).

These tools are great for diagnosis of important high level problems. Important because they directly affect your customers, but high level because they arent the causes, only the effects. This is a good place to start to discover the minor faults and incorrect settings that cause inconsistancy & therefore SEO penalisation and start to get a cleaner faster more reliable website.