How to manage service delivery in Microsoft Exchange?
Service Delivery for Exchange is more than complex because of the Role architecture.
Even with multiple role servers that save hardware costs, how can you ensure that because an Exchange VM is up, with Client Access Server, Hub and Mailbox roles on it, the service delivered by these three roles is also up?
The uptime of the Exchange server has nothing to do with the ability for a mailbox role to be replicated, for a mailbox database to be available or with the ability for the Hub to proceed with all the mail queue, or finally for the CAS to allow connections through different protocols that you have configured.
Exchange is a collection of services provided by several applications and therefore the only way to measure its availability is to measure the one delivered by each of these “applications”.
To manage the service delivery you need to measure different services and prevent any of them from impacting the user experience.
GSX Monitor & Analyzer main features
Measuring the service availability
Things are simple when you have a simple environment: One multiple role servers, no Database Availability Group, no Client Access Server Array.
In this case you can simply assume that the availability of each service is directly linked to the role availability and if one role is down, the global service provided by Exchange is down. This is the case for a multirole server as well as for a single role collection of servers.
Unfortunately, this infrastructure is not very common on the market place. You would have at least two Client Access Server to enable high availability of the protocols, two mailbox servers to enable high availability of your databases, and sometimes multiple Hub servers depending on the number of sites you have.
Well, things are becoming more and more complex.
That is why the best indicator that provides you with valuable metrics is the end user experience on any of your Exchange role server mixed with the availability of your mailbox databases.
How GSX can help you?
As you can see in the roles’ pages, GSX considers the service delivered by each role at multiple levels. At the server level, at the database level and at the “cluster” level (Database Availability Group and Client Access Server Array).
First service: The ability for a user to connect to its Mailbox
Regarding the service delivered by the Client Access Server, GSX measures and reports on end user tests made via Synthetic Transactions that simulate a real action of a user, reports a latency time for you to be alerted before it impacts users and calculates the availability of the service from a user perspective, at the Client Access Server level:
- Full Synchronization with ActiveSync Mobile
- Full connection to a mailbox with OWA
- Full connection to mailbox with POP3/IMAP
- Full utilization of the Autodiscover checking the configuration of availability Service, Unified Messaging Service, Outlook Anywhere, Offline Address Book.
And at the Client Access Server Array Level:
- Full measure of the service provides by the CAS Array
- Load balancing and Network Load Balancing usage
Hence GSX provides valuable metrics of the availability of the service delivered by the Client Access Server. See the Client Access Server and Client Acces Server Array section for more information.
The Client Access Server service measures the ability for user to connect to their mailbox. GSX collects historical data for you to trend this availability against your infrastructure.
Second service: The ability to send and receive mail
GSX simulates user activities to give you a true availability of the service provided by the Hub. First, at the server level:
- Full check of the mail queues
- Full control on messages in the multiple queues
- Full control on mail flow with historical Data
And at the organization level:
- Full end to end test on sending / receiving message internally
- Full end to end test on sending / receiving mail to internet
- Full end to end test on sending / receiving to a mobile device
Hence GSX provides true measure of the availability of the service delivered by the Hub infrastructure. See the Hub section for more information.
Third service: The ability to see and manipulate data in the mailbox
GSX checks constantly the availability and the high availability of the mailbox databases, collecting historical data to give the ability to trend and forecast any of those statistics. It also checks the possibility for any user to access and manipulate its mailbox.
- Full connection to mailbox and use of the data
- Full MAPI, LDAP, DS Access checks
- Full checks on database availability at the server and at the DAG level
- Full set of statistics on database usage
Hence GSX provides true measure of the availability of the service delivered by the mailbox Role server and the mailbox databases infrastructure.
See the mailbox role server and mailbox database management section for more information.
Reporting on Service Availability
Measuring is good. Have graphical interpretation is better. GSX Analyzer collects all these measurement and allows administrator and IT Manager to use built-in reports and customize it if needed. Being able to forecast your service availability data in GSX Analyzer just take one click.
No script anymore to collect data, no complex extract from SQL or endless aggregation in Excel. Everything is directly available in GSX Analyzer.
GSX Analyzer draws the availability of these three main services provided by the Exchange infrastructure for you to report easily.
Measuring SLA, comparing SLA against a contract value or just building high level reports on availability as well as very technical one for the technical team is a matter of minutes.
Identify bottlenecks and capacity issue
Managing the Service Availability is not only measuring the final availability and report on it.
It is also measuring the performance of each point, identifying any bottleneck and capacity issue to anticipate issues before they impact the end users. The goal is to tend to 100% availability.
For that the performance is linked to the reporting part that shows in seconds the point of failure of your Infrastructure.
GSX Analyzer reports on all the tests and all the statistics collected by GSX Monitor to compare instantly the health of your servers and detect where you have to focus your efforts. It compares the availability of each component of your infrastructure, mailbox server with each other, CAS server with each other, Hub, mailbox databases, etc to quickly identify bottleneck and capacity issues.
The Forecasting feature helps you define your capacity planning, detecting when you’ll be out of storage, when the mailbox database will kill your free space or even when the RAM won’t be enough on your server.
Capacity isn’t only storage; it is the ability for the infrastructure to support the users demand. Any of the GSX statistics can be drawn and compared for your capacity planning.
Taking actions before it impacts the users
GSX collects and measures performance statistics of your environment. But it also alerts you when this performance is decreasing to take action before any user impacts.
Performance like connection to the mailbox, usage of the mailbox, ability to send, receive mail, etc. is tracked and subject to threshold for proactive action. The GSX alert system is as simple as powerful providing multiple ways to be alerted including escalation process, time zone and multiple profiles.
Make admin life easier
Managing service availability is a complex action in Exchange 2010 due to the multiple services Exchange provided through the roles. And knowing how to balance the CAS service with the mailbox, mailbox database, Outlook Web Services, Outlook anywhere, Hub, etc. will always be complex.
GSX helps you measuring the true availability provided by the collection of service, allowing easy comparison and easy report on service availability.