Monitor Host Instances in Clustered/High availability environment
A BizTalk Server Host is a logical container that contains the BizTalk Server runtime processes in which you deploy items such as adapter handlers, receive locations (including pipelines), and orchestrations. The host instance is the runtime process where the message processing, receiving, and transmitting occurs. Obviously, it is very important to ensure your host instances are running in your BizTalk environment. BizTalk360 already made it easy for its customers by introducing the Host Instance Monitoring functionality. In BizTalk360, you can monitor the BizTalk host instances by setting the Expected State. If the expected state doesn’t match with the current state of the host instance, you will get notified.
Out of the box, BizTalk360 allows users to monitor the Clustered Host Instances using the “AtLeastOneActive” state.
If this state is assigned as the Expected State, then the monitoring service will verify that across all the instances for the respective clustered host, at least one host instance is active, guaranteeing that the server is running, and no downtime happened for that host/server. To know more about Clustered Host Instance monitoring, follow this blog link.
In the upcoming BizTalk360 version 9.0 Phase 2, we have extended this core functionality to monitor the host instances for high availability environment, as per the customer feedback and voting as below.
BizTalk Server High availability
The following figure shows a BizTalk Server deployment that provides high availability for the receiving host by having two host instances on different computers. Note, that in this figure the processing and sending host is not highly available, because there is only one host instance processing the BizTalk items assigned to the host.
How BizTalk360 helps to monitor the Host Instances in clustered/High availability Environments?
In a business scenario, if you have two BizTalk servers and they both have a Host Instance on them called “FTP”, both Host Instances will trigger a connection to this FTP server. If a file exists on the FTP server that matches your Receive Location’s file mask, both of these host instances will attempt to retrieve a copy of this file, as shown below.
To avoid processing duplicate messages from the high availability, organizations may use the Windows Clustering. This will guarantee that only one host instance is running at a time.
In the above cases, any one of the BizTalk servers needs to be in the active state to receive the file, if both the server is down, then the file will not get received. With BizTalk360, a user can monitor both the host instances in different servers, by setting up the “AtleastOneActive” State as shown below.
In High Availability environments, users can have all the host instances to be Active. For example, if you have 3 BizTalk Host Instances, in multiple BizTalk servers, used for different purposes like receiving, processing and sending the files.
You can set up monitoring for the above scenario in BizTalk360, by setting the expected state to Started for all the three Host Instances as you can see below. When any of the host instances goes down, you will be notified.
Setting up the “AtleastOneActive” state for Host Instances
The “AtleastOneActive” state is mainly used to monitor the Active/Passive clustered and Non-clustered host instance monitoring. The user can set the “AtleastOneActive” state only for clustered host instances and non-clustered host instances in high availability environment.
Setting the “AtleastOneActive” state will monitor across all the configured host instances of the same host and check if any one of the host instances is active. If so, the monitoring state will be healthy.
If all the configured host instances are down or stopped state, then the host instance monitoring will turn to Critical state.
The monitoring status of the host instances will get reflected in the BizTalk360 Monitoring dashboard and the same will be notified in detail in the alert email.
How does AutoCorrect functionality work for the AtleastOneActive state of the host Instance monitoring?
Do you feel frustrated to manually start host instances, when all the host instances become down in clustered/High availability environment?.
BizTalk360 comes up with another solution with its Auto Healing capability.
For Example, if you have 2 BizTalk host instances running under Clustered/high availability setup, from 2 host instances any one instance needs to be in the active state for performing file transactions. If both the host instances are down, then you need to start the host instances manually. This requires continuous monitoring, which can be handled by BizTalk360 itself, by configuring the AutoCorrect functionality to the clustered or non-clustered host instances.
In the above screenshot, the auto-correct functionality is enabled for 2 host instances in different BizTalk servers BTS1 and BTS2. If both the host instances are down, from next monitoring cycle BizTalk360 will try to auto-heal any one of the host instances from these servers. On the first attempt, it will try to start the host instance in BTS1 server. If the instance is started successfully then it skips auto-healing the next server BTS2. In case BizTalk360 fails to start the instance in BTS1, then it tries to start the host instance on the server BTS2. BizTalk360 will try to start any one of the host instances until it reaches the maximum attempt count. You can also reset the auto-correct by setting the reset interval. To know more about the auto-correct feature click here.
Monitoring the Host Instances in Clustered and High availability environment is a must for a healthy BizTalk environment. BizTalk360 monitors those host instances in an easy and simple manner. It also manages the host instances and ensures they are up and running, by using the auto-correct functionality. The upcoming version of BizTalk360 will ease your way of monitoring the host instances in clustered and high availability setup.