Application rollouts, major upgrades or widespread patching can wreak havoc on network performance, but a new survey finds that even minor day-to-day infrastructure changes can send performance plummeting.

The survey, conducted by King Research and Usermonitoring.org, and co-sponsored by application performance management vendor Coradiant, polled more than 200 IT professionals about the impact change has on web-enabled applications.

The results show application patches (nearly 40%) and day-to-day infrastructure changes (about 37%) caused the majority of unplanned outages. While respondents indicated they are aware large planned changes could cause outages, the research firm finds that smaller changes seem to have the same impact as larger alterations.

"Slightly more outages were caused by application patches and infrastructure changes than by either the release of a major application upgrade or the rollout of a new Web application," the report said, indicating “minor changes are just as risky as major application releases".

It also found that, while network managers would like to spot the performance problem before users and customers notice the service degradation, 72% of respondents find out about poorly performance applications from user reports and complaints.

About half the survey participants use service-level measurement tools to validate when a problem has been corrected, and about one-fourth said: "They assume problems are corrected if they don't hear further complaints."

Some 37% of survey respondents reported that they rely heavily on expert opinion when it comes to diagnosing the cause of performance problems, while 22% use internally developed diagnostics tools and 21% use commercially available tools.

"IT continues to rely heavily on end users to validate that problems have been corrected with fewer participants reporting the use of measurement tools," the report concluded.