Sign up & instance setup · BYOF: identity, access and monitoring
Set up live Fabric capacity monitoring with Eventstream
Create the two capacity sources, publish their destination and prove that Emplyr is collecting both event types.
BYOF setup checklist · Task 4 of 5
On this page
Before you start
- Use the correct instance Admin membership and a tested BYOF Fabric connection. Have the exact capacity name, ID and region. Activate the approved capacity in Connect live usage before the final instance verification.
- In Fabric, the person building this needs Contributor or higher in a Fabric/trial workspace and capacity-admin authority for the monitored capacity. Place the Eventstream in the same region as that capacity. These are the human operator’s rights, separate from the app’s permissions.
- Arrange the persistent event collector with your Emplyr deployment administrator. Creating a Fabric destination does not start that service. For Managed instances the platform team owns the shared stream; customers use its displayed status.
Keep these checks separate
On smaller screens, scroll the table sideways to see all columns.
| Checkpoint | What it proves |
|---|---|
| Capacity activated | Commercial access for this BYOF instance; not source publication. |
| Destination saved securely | The destination configuration is saved/tested; not collection. |
| Both received statuses | The running collector has received the required source events. |
| Verify events & continue | The instance’s live readiness check succeeds. |
| Metrics app / history import | Optional historical context; cannot replace live verification. |
Shared streams and workspace changes
A destination can carry several capacities when each has both required sources and approved access. A new content workspace on an already monitored capacity does not by itself require another stream; check its access and capacity placement. Keep region requirements in mind when designing shared streams.
Follow these steps
Screenshots use fictional training data. Names, dates and amounts are examples. Select an image to view it at full size.
- In Emplyr setup → Connect live usage, find the approved activated capacity and choose Set up stream (or Manage stream). For an existing instance use Admin / Global → Fabric → Capacities. Check the wizard’s capacity name/ID before opening Fabric.
- In Prepare, open the agreed Fabric workspace → New item, search Eventstream, enter a meaningful name such as emplyr-operations-usage and choose Create. Confirm the workspace’s capacity region matches the monitored capacity.
- In the Eventstream editor choose Connect data sources (or Add source → Connect data sources). Search for Capacity operation events, choose Connect and select Microsoft.Fabric.CapacityOperationEvents.Operation. Scope it to the exact monitored capacity, name the source and complete Next → Add.
- Add a second source the same way: Capacity overview events. Select Microsoft.Fabric.Capacity.Summary and the same capacity, name the source and add it. Capacity state-change events alone do not supply the required summary. Keep both original event envelopes intact.
- In Edit mode choose Add destination → Custom endpoint, give it a name and Save. Connect the default stream carrying both sources to that destination, then Publish. Avoid transformations that drop fields or replace the original JSON envelope.
- Switch to the published Live view and select the destination. Open its Details/Keys area → Event Hub → SAS Key Authentication. Use the destination’s Event Hubs-compatible connection string, not a source endpoint, REST URL or Kafka credential.
- Return to Emplyr → Add sources and tick I added both sources for this capacity. In Connect destination paste the Event Hubs connection string. Supply Event hub name if the string has no EntityPath; set Consumer group to the one actually used by the collector ($Default is the form’s default). Select Save & verify.

After publishing the Fabric destination, save its connection details here. Check saved stream uses the existing protected destination; verification must also confirm live event receipt. - Open Verify collection. Ask the deployment administrator to confirm the persistent collector is running against this destination and consumer group. Generate real work on the monitored capacity, for example a notebook run or semantic-model refresh. The wizard rechecks roughly every 15 seconds.
- Check all three statuses: Destination saved securely, Overview events received and Operation events received. Finish setup can be available while events are still awaited; record delivery success only after both received statuses appear.

Check activation and receipt of both event types before Verify events & continue. Optional usage history is a later step. - For new-instance setup, close the stream wizard and choose Verify events & continue in Connect live usage. Resolve any failed readiness checks. For an existing ready instance, open Capacity usage, select the same capacity/date range and confirm incoming activity. Repeat for each intended capacity.
Worked example
NovaNest monitors its Operations F-capacity from an Eventstream workspace in the same region. Its capacity administrator adds Operation and Summary sources scoped to that capacity and publishes a Custom endpoint destination. After the deployment administrator confirms collection, a test notebook produces activity. Both received indicators turn successful and the instance administrator completes Verify events & continue.
Microsoft setup references
What you should see
Both sources are published for the correct capacity, their destination is saved securely, and Emplyr confirms recent overview and operation events. New-instance live verification succeeds independently of optional history.
If something goes wrong
Save & verify succeeds but no events arrive
That action checks destination connectivity/metadata; it does not prove upstream publication, consumer-group existence or event delivery. Recheck the published path and both source scopes, then have the deployment administrator check collector health and its actual consumer group. Refreshing a browser/report does not run the collector.
Overview events arrive but operations do not
Check the operation event type and capacity scope, the source creator’s capacity-admin rights and same-region placement. Run actual workload activity and wait for delivery; an idle capacity may not produce new operations.
Only one of several capacities is ready
Both event types must reach Emplyr for each intended capacity. Add both sources for the missing capacity. Reuse a saved destination only when that destination really carries those sources too.
The wizard says Managed by Emplyr
The platform team maintains that stream. Supply your organisation’s contact with the safe instance/capacity reference and missing-event status; customers do not replace Managed credentials.