Feedback
Processing Overview
This dashboard gives a consolidated, at-a-glance view of Relativity Processing activity and infrastructure health. It reports how many inventory, discovery, and publish jobs have completed, how many jobs are currently in progress across processing and imaging, and the health of the queue manager and worker hosts. It helps administrators confirm that processing is running, spot stalled or backed-up work, and verify processing health after installations, upgrades, or configuration changes.
Availability
The Processing Overview dashboard is available in Relativity Server 2026 and later. Its visualizations depend on processing instrumentation shipped with the server and the Invariant processing components. If the dashboard loads empty, the environment does not yet include the required instrumentation, or no processing activity or heartbeats were reported recently.
Prerequisites
- Environment Watch is installed at a release that includes the processing jobs status metrics (REL-1381296), which provide the Invariant SQL query receiver used for the job-count tiles, and the Elastic/Kibana stack is receiving telemetry from the target Relativity Server.
- The Relativity Server is version 2026+ so that processing OpenTelemetry (OTEL) metrics and health checks are emitted.
- The Environment Watch agent is running on the Invariant Queue Manager server and has access to the Invariant database. It uses a dedicated SQL Server connection to the Invariant database (supplied to the OTEL collector as
OTEL_INVARIANT_SQL_SERVER_CONN_STRING) to collect job counts. - Processing jobs have run within the selected time range, and queue manager and worker heartbeats have been received within the last 30 minutes.
Navigation
The Processing Overview is part of the Processing feature domain. A tab strip at the top of the dashboard lets you switch between the related views:
- Processing Overview
- Platform Health
- Inventory & Discovery
A left-hand navigation panel provides quick links to other Environment Watch feature domains (Platform, Applications, Logs/Searches/Alerts).
Filters and Time Range
- Time range - Use the time picker in the top-right to scope the panels. The dashboard does not save a fixed time range, so it opens with your Kibana default (or last-used) range. The Queue Manager Health and Worker Health tiles ignore the time picker and always use the last 30 minutes of heartbeats.
- KQL filter bar - Use the query bar to narrow results with Kibana Query Language and add ad-hoc filters.
Job Totals and Server Availability
The tiles across the top of the dashboard summarize processing job totals and server availability:
- No Of Inventory Jobs Completed - Total inventory jobs that have reached the completed state.
- No Of Discovery Jobs Completed - Total discovery jobs that have reached the completed state.
- No Of Publish Jobs Completed - Total publish jobs that have reached the completed state.
- Queue Manager Health - Status badge for the queue manager hosts, with an X / Y Active line beneath it. A host is active when its latest heartbeat is within the last 20 minutes. The badge shows one of three states:
- Healthy - Every host seen has sent a heartbeat within the last 20 minutes.
- Unhealthy - At least one host has had no heartbeat for more than 20 minutes.
- N/A - No hosts were seen in the last 30 minutes, or the results are incomplete.
- Worker Health - The same status badge and X / Y Active line for worker hosts.
Refresh the dashboard to fetch new heartbeats.
The completed-job totals are point-in-time counts of rows in the Invariant Jobs table, read at each collection interval (every 60 seconds), and each tile shows the most recently reported value. They can decrease when Invariant purges job rows.
In Progress Jobs
The In Progress Jobs donut shows the peak number of jobs in progress, by job type, over the selected time range. To see the current state, use a short time range such as the last 15 minutes. Job types:
- Inventory - Inventory jobs in progress.
- Discovery - Discovery jobs in progress.
- Publish - Publish jobs in progress.
- Imaging - Native and basic imaging jobs currently running (the two imaging job types are combined).
Use this view to see which kind of processing work is currently dominating the environment. A large or unchanging in-progress count combined with no new completed jobs can indicate stalled work or unavailable workers; check the Platform Health tab next.
Metrics Reference
OTEL metrics
Job counts are collected from the Invariant database by the Invariant Queue Manager collector. Each processing job type emits the same set of counts, where <type> is discovery, inventory, or publish.
| OTEL metric | Description |
|---|---|
relsvr.processing.<type>.jobs.completed_count
|
Jobs of the type that have completed. |
relsvr.processing.<type>.jobs.in_progress_count
|
Jobs of the type that are currently in progress. |
relsvr.imaging.basic.jobs.running_count
|
Basic imaging jobs currently running. |
relsvr.imaging.native.jobs.running_count
|
Native imaging jobs currently running. |
Server availability is derived from the Invariant heartbeat (running health check) logs:
Log name (labels.name) |
Description |
|---|---|
relsvr.invariant.queuemanager.runninghealthcheck
|
Heartbeat emitted by a queue manager host. |
relsvr.invariant.worker.runninghealthcheck
|
Heartbeat emitted by a worker host. |
Panel-to-metric mapping
| Panel | Type | Metric(s) / source | Key fields |
|---|---|---|---|
| No Of Inventory Jobs Completed | Metric | relsvr.processing.inventory.jobs.completed_count
|
last value |
| No Of Discovery Jobs Completed | Metric | relsvr.processing.discovery.jobs.completed_count
|
last value |
| No Of Publish Jobs Completed | Metric | relsvr.processing.publish.jobs.completed_count
|
last value |
| Queue Manager Health | Status badge | relsvr.invariant.queuemanager.runninghealthcheck
|
unique labels.host_name, latest heartbeat time |
| Worker Health | Status badge | relsvr.invariant.worker.runninghealthcheck
|
unique labels.host_name, latest heartbeat time |
| In Progress Jobs | Donut | relsvr.processing.inventory.jobs.in_progress_count, relsvr.processing.discovery.jobs.in_progress_count, relsvr.processing.publish.jobs.in_progress_count, relsvr.imaging.basic.jobs.running_count, relsvr.imaging.native.jobs.running_count |
max value per job type |
Use Cases
| Use Case | Description |
|---|---|
| Confirm processing is running | Check that completed job totals are advancing and that jobs show as in progress when work is submitted. |
| Verify infrastructure availability | Use the Queue Manager Health and Worker Health tiles to confirm the hosts are Healthy and the expected number are active. |
| Detect stalled work | Compare a persistent in-progress count against flat completed totals and inactive workers to identify stuck or starved jobs. |
| Understand workload mix | Use the In Progress Jobs donut to see how inventory, discovery, publish, and imaging work is distributed. |
| Validate post-install/upgrade health | After a deployment or configuration change, confirm queue managers and workers report heartbeats and jobs complete. |
Associated Alerts
No processing-specific Environment Watch alerts are currently associated with this dashboard. This section will be updated as processing alerts are published.
