A few months ago I spent some time with a customer that was suffering from periodic SQL Server outages. In most of these cases, the database servers were not actually going down, but applications were having difficulty connecting to them due to excessive resource utilization. This particular company provided a service, direct to consumers that relied on the availability of their data. Any outage that impacted the availability of their service would at a minimum result in poor customer satisfaction, and could have long term ramifications that included a loss of customers and revenue. To make this matter worse, at a high profile meeting with the customers IT leadership team, an outage was being discussed that the DBA Manager had not been made aware of. His team had quickly addressed the outage and brought the system back online, however this particular manager was unaware of the problem, and had no way to view historical availability over time to demonstrate whether they were meeting the SLA's that the business had defined for them.
In order to solve these specific challenges, Foglight was considered to:
- Proactive alerting to be notified about performance issues before database availability is impacted.
- Rapid diagnostic capabilities to pinpoint the problems root case, and fix the problem before it brings the system down.
- Provide enterprise visibility across the customers hundreds of SQL Server instances, to quickly view which servers are down.
- Historical reporting to view availability SLA attainment over time.
During the evaluation of the software, the customer quickly realized the power of Foglight for SQL Server's performance monitoring tool and it's ability to quickly alert on, and diagnose performance issues, however they were struggling to provide the exact type of enterprise visibility that this one particular DBA manager was after. The "out of the box" Databases dashboard did indicate that a SQL Server Instance was down, however the customer wanted it to "jump off the page".
Foglight custom dashboarding to the rescue!!! After taking a few requirements from the customer, we were able to provide them the dashboard below, which their management could use to track Enterprise SQL Server Availability, both real time, and historically. Notice the "red" server indicating that the server is down as well as the second table showing that the server has only been available for 41% of the time over the defined time range.
For those interested in this dashboard, and several others, we've made it available in the latest release of the Database Expansion Pack (version 5.0).