Nia CoreMonitoring and control
The Agents page
Lists every agent paired to this workspace: its online status (based on when it last checked in), its version, and its jobs — both published from a Nia Core workflow and any created directly on the agent machine (shown marked "local", with status and run history visible but no remote controls, since those stay fully agent-owned). For a published job, you can also see the version the platform wants it running versus the version the agent has actually applied, and why, if it rejected an update.
On the agent machine itself, nia-agent status shows the same information from that side: how long the agent has been running, whether it's paired, when it last checked in, and the health of each job.
Job health
| Status | Meaning | What to do |
|---|---|---|
| ok | Running on schedule, no unresolved failure. | Nothing. |
| failing | The most recent run(s) failed, but the agent is still retrying automatically on schedule. | Check the error class for that run; most clear up on their own. |
| paused | The agent stopped running this job automatically. | Fix the underlying issue, then Resume. |
Last run result
| Result | Meaning |
|---|---|
| completed | The run finished and sent its rows successfully. |
| failed | The run did not complete — see its error class below. |
| skipped | A scheduled run was skipped — either the job is paused, or the previous run of the same job was still in progress. Not an error. |
Error classes and what to do
"Pauses" means the job stops running until you Resume it; everything else keeps retrying on its normal schedule (some get fast automatic retries too).
| Error class | What it means | Pauses? | What to do |
|---|---|---|---|
| config | The job's own setup is invalid — e.g. the destination rejected the credentials, or a required setting is missing. | Yes | Fix the destination credentials/settings, Test, then Resume. |
| schemaDrift | The destination's columns have changed since this job was set up. | Yes | Review what changed, update the mapping, Test, then Resume. |
| typeMismatch | A value didn't fit the type the destination expects for that column. | Yes | Check the source column and mapping, fix the mismatch, then Resume. |
| nullKey | A row had a missing or blank key value, and the job is set to stop rather than skip it. | Yes | Clean up the data, or allow skipping null-key rows, then Resume. |
| massDelete | A sync would delete a larger share of destination rows than allowed. | Yes | Confirm the deletion is expected, raise the limit or use "Allow one large delete," then Resume. |
| transient | A network problem, or a temporary error from the destination. | No — retries at 1, 5, and 15 minutes | Usually nothing; if it persists, check connectivity. |
| diskSpace | The agent's local staging area is low on free disk space. | No — retries at 1, 5, and 15 minutes | Free up disk space on the agent machine. |
| lockTaken | Another sync against the same database server was already running. | No | Nothing — it tries again next scheduled run. |
| rejected | The destination flatly rejected the request (not a temporary one). | No | Check the destination settings; contact support with the job ID and timestamp if unclear. |
| mismatch | The destination reported a different row count than what was sent. | No | Usually transient; contact support if it recurs for the same job. |
| emptyReplace | A full-reload run found zero rows, and the job isn't configured to allow that. | No | Confirm the source has data, or allow empty replace once, then re-run. |
| aborted | The run was stopped mid-way — e.g. the agent restarted. Not a real failure. | No | Nothing — it runs again on schedule. |
| other | Anything outside the classes above. | No | Check the error message; contact support if it recurs. |
Controls
- Pause / Resume — stop or restart a job's schedule.
- Run now — run once immediately, with optional one-time parameter overrides.
- Force full reload — run once, ignoring the normal mode, resending everything.