Below you'll find answers to the questions we get asked the most about the Exalate app.
When connecting to the Exalate network, you are able to synchronize as long as your remote partner is fully paying for the synchronization using the Exalate network license.
You can install the app unlicensed and keep the License Key field empty in this case.
Exalate uses a Groovy-based scripting engine to implement filtering, transformation, and mapping. Even though scripting like that is more complex than a drag and drop interface, it is the only way that makes even the most complex mapping possible to implement. See how the most complex use cases can be implemented using the synchronization processors.
There are 2 different situations to be considered
Any new issue brought under sync will be synchronized normally
All issues under sync will require a 'calibration phase' which allows to bring the synchronization back into shape.
Every replica of an issue is stored on 2 systems. A replica is the result of the transformation of an issue into a hub issue (through the data filter) This replica is stored:
on the source, tracker to detect if relevant changes have been made since the last synchronization
in the destination, a tracker to provide the content of the previous version.
Yes - the synchronization panel will detail out all information relevant to the synchronization.
Not yet - we are working on it. When staging exalate 1.3 or earlier, make sure that the synchronization engine is not triggered by deactivating all instances.
When you delete an issue Exalate consider it as a synchronization event. It triggers a number of events which ensure that future synchronization events are not processed anymore. The synchronization panel will indicate that the remote issue has been removed.
Not at all - every application administrator can define how 'hubissues' are translated to the local context.
No - the intermediate layer is taking care to transport whatever information the originating tracker wants to share. It is up to the other side to take it into account.
Yes. Right now, Exalate is compatible with JIRA Server, JIRA Cloud, HP QC ALM, GitHub, and Zendesk. We are working on the integration with other issue trackers.
There are no limits to sending information using the customKeys property of a replica. It allows sending any arbitrary object which will get serialized at one end and deserialized at the other.
Synchronization is unlimited between supported platforms as long as Exalate is installed on each instance. We provide 2 licensing models:
- Instance licensing - the model, where each instance requires a valid license to synchronize.
- Network licensing - the model where only one side pays for the synchronization between instances.
Exalate has a built-in scripting engine to implement almost any use case in a flexible way. Check our use case examples for more details.
Optionally. The application administrator can configure how links to the remote issue are represented. Either as JIRA issue link, as a weblink in the synchronization panel or not at all.
Exalate is already in use since 2014 at different enterprises. The Exalate has been configured as a single-threaded application which processes the synchronization transactions one at a time. This has 2 consequences
- A minimal system load, as the exalate can be compared to a single user who is working hard to update/retrieve information
- It takes time to process transactions (because of the single-threaded nature) and the product is not meant to do large migrations
One deployment has on average 12000 synchronizations/month without a significant impact on the overall instance performance
From JIRA to HP: Changes are processed almost immediately, depending on the number of outstanding synchronization requests.
From HP to JIRA: The exalate application is querying the HP application every 5 seconds for any new change, and processes these (using a single thread). If the number of changes requires more processing time than 5 seconds, the next cycle is run immediately.
From JIRA to JIRA: Changes are processed almost immediately, depending on the number of outstanding synchronization requests.
In 3.2.1 release subscriptions page was hidden. But the page is still there. You can access it using this kind of link :
Thanks to the distributed model, it is possible to adapt the common hub issue to the local context. For instance, if your local workflow evolves, you will be able to change the create and change processors to take into account the new configuration.
When an instance requests another instance to send over the content of an attachment, the other instance will respond with a hash key calculated from the content of the attachment. The same hash key is calculated on the receiving end to ensure that the received content matches the original attachment. The download will be retried once to avoid transmission errors. Failing this second attempt an error is raised and the administrator is notified.
Exalate is based on a distributed model where each application administrator configures what information can be send and how incoming messages must be handled. All messages exchanged between the nodes are based on a common defined 'hubissue' which carries all information one application administrator wants to communicate to the other side. The processors will use this information to apply the changes to the local issues.
Absolutely. The synchronization panel shows all relevant details - even if the remote issue has been removed.
Yes. Thanks to the groovy goodness, one can parse the newly added comments and trigger a resolve transition
A scenario we encounter regularly is the service desk use case where tickets raised by a customer need to be raised internally in different projects - depending on the set of properties. One use case was particularly interesting. The target project depended on 3 different properties resulting in 200+ potential permutations. This mapping is stored in a single database table.
The way we solved it involved finding the right projectkey in the database table and use it to raise the issue in the right project. Check following example showing how such configuration can be set up using Exalate.
The synchronization engine is processing sync events sequentially (for the moment - November 26. 2016). Processing 10000 issues in 100 different projects might take a couple of hours. We have been synchronizing 10k+ issues as part of our performance tests and we are looking at increasing the processing speed to meet the challenge that one of our prospects is facing.
Yes. Each integration point can evolve as long as the hub issue model and Exalate API is respected.
This happens immediately. Exalate is based on a fully event-driven engine Whenever a user changes some information (and that information needs to be sent), a sync event is generated containing a snapshot of the modified issue - ready to be sent to the remote instance. When 10 issues are changed, or 1 issue 10 times at the same moment (it does happen) - 10 events are generated.
The replication layer in the exalate architecture will transmit each sync event to the remote instance, triggering a sync request, which is processed at its own pace.
Handling a full synchronization event takes a couple of seconds (give/take the size of the information to be transmitted)
Whenever an error occurs, a detailed overview (= stack trace) leading to the problem is raised. In case this is not sufficient, one might create a support.zip and send it to email@example.com for further processing.
Since Exalate is an app for different deployments of Jira (Server, Cloud, Data Center) and other platforms, we do not attempt to ensure that our application is compatible with all third-party reverse-proxy solutions that might be set up on the Jira host.
However, we've tested it with nginx, and we can state that Exalate is compatible with nginx as a reverse-proxy server.
The group of exalate administrators will be notified in 2 different ways
by email with the details of the error notification
In JIRA - by using 'In-JIRA' notifications - popping up a warning about the blocking error
Also, errors are generated at 4 different levels.
As an example - when the proxy user is not allowed to modify the local issue due to issue security
Typical example - when there is an error in the processor scripts, or when there is a permission problem to apply changes to an issue
A connection problem
Bugs in the synchronization layer.
Whenever a user changes something in an issue, Jira will notify the Exalate about the change by generating an update event. If the issue is under an update gets synchronized.
However, changing the key does not trigger such an event.
To synchronize a key change, the issue under sync needs to be touched by modifying something or by adding a comment in order to synchronize that information change.
Additional considerations when changing the project key
- Review the Create Processor to make sure that the old project key is not hardcoded.
- Review the triggers and update the JQL if this is relevant
Yes, it's possible to sync multiple projects with one connection.
Basically, it's all about the configuration of the Sync Rules. Once that you establish a connection between instances you can specify which data would you like to send to the other side and how to interpret incoming sync data.
The Data Filter of the sending side interacts with Create and Change Processor of the receiving side.
For more details, please check the synchronization processors.
Versions and pre-requisites
Currently, Exalate supports JIRA Core, JIRA Service Desk, JIRA Data Center, JIRA Software, JIRA Cloud, HP ALM. We are working on some other issue trackers to be compiled with Exalate
All editions of JIRA (Core, Software, Service Desk) and platforms (Server, Datacenter and Cloud). Server and datacenter supported from JIRA 6.3 and up.
We are validating against HP ALM 12.02, and 12.50.
For now only HP ALM defects could be synchronized.
For JIRA Server: Exalate is an Atlassian add-on, which can be installed as any other add-on. Synchronization information is stored in the JIRA database. The memory footprint is limited and should not require any modification from the existing setting. In case of doubt, adding 128MB to your jvm settings should be sufficient.
You will be required to configure a single JIRA user, which will be used to access the issues to be synchronised
Exalate also requires a SMTP server to raise notifications in the case of errors. For this SMTP a hostname, userid and password are to be provided in case outgoing mail requires authentication
For JIRA Cloud: Exalate is an Atlassian add-on, which can be installed as any other add-on. Synchronization information is stored in the JIRA database.
For HP ALM: Exalate is delivered as an external application which interfaces with HP through the REST API. You will need following:
A host which can run the application
The application is delivered either as a RPM (centos) or as a docker image (many platforms)
In the case of an RPM, a postgresql 9.4 database needs to be deployed on the host
The exalate application needs to have access on the HP ALM application (for reading and writing)
In 4.0 version we've introduced Connection and Invitation concept that includes changes to the User Interface.
The configuration is simplified now. What we used to call Instances and Relations is now merged into Connections.
The Connection includes information from the Instance and a Relation, based on that Instance.
The distributed architecture of the synchronization engine allows to upgrade either end as appropriate. The common hub issue model ensures that the synchronization information can be translated to the local context.
For customers who already use the Exalate Subscription and now wish to upgrade to 2.8 and higher. We've decided to give a simpler licensing experience and introduced Atlassian Marketplace licensing support. Find more details.
All you need to know before installing the Exalate app for HP QC/ALM could be found here.
Exalate can be used with JIRA 6.3 and higher. Due to incompatible API's between JIRA 6 and JIRA 7, it is being published as 2 different distributions. For administrators installing using the universal plugin manager (UPM) there is no problem, as the UPM will point to the right distribution - which is pretty convenient. Still - there are a lot of instances which are not permitted to reach out to the marketplace.
The marketplace is hiding versions which are compatible with older JIRA releases. Find more details.