Skip to content

Troubleshooting

Common errors during install, collection, or dashboarding — and what to do about them.

Install-time

The term 'Invoke-DbaQuery' is not recognized

The dbatools module wasn't importable in the runspace. Check DbaToolsFolderPath points at a folder whose layout is:

text DbaToolsFolderPath\ <version>\ dbatools.psd1 dbatools.psm1 ...

Try:

powershell Save-Module dbatools -Path 'D:\Downloads\dbatools' -Repository PSGallery

then pass -DbaToolsFolderPath 'D:\Downloads\dbatools'.

Access is denied from PSRemoting

The wrapper can't reach the monitored host as WindowsCredential. Check, on the host:

powershell Enable-PSRemoting -Force -SkipNetworkProfileCheck Set-ItemProperty HKLM:\SOFTWARE\Microsoft\Windows\CurrentVersion\Policies\System ` -Name LocalAccountTokenFilterPolicy -Value 1

…and on the deployer:

powershell Set-Item WSMAN:\Localhost\Client\TrustedHosts -Value 'MyHost.lab.com' -Concatenate -Force

If the host is cross-domain, pass a domain-qualified credential (DOMAIN\svc_sqlmonitor) or use CredSSP.

Linked server creation failed: Msg 15457

Means you're creating a linked server with SQL auth but the target instance rejects the password. Re-check:

  • grafana SID matches between inventory and monitored instance (step 56__GrafanaLogin with -DropCreatedLogin $true to force-sync).
  • TrustServerCertificate or a valid cert on the target — TLS 1.2+ is mandatory on modern Windows SQL clients.

Invalid column name 'X' during 2__AllDatabaseObjects

Schema drift from a partial previous run. Rerun step 2 with an elevated error level:

powershell -OnlySteps '2__AllDatabaseObjects' -Verbose -ReturnInlineErrorMessage $true

If a specific table is the culprit and you can afford to lose its data:

sql DROP TABLE dbo.<table_name>; -- then re-run step 2

The XEvent session 'XEventMetrics' already exists

Either drop and re-create or skip:

sql IF EXISTS (SELECT 1 FROM sys.server_event_sessions WHERE name = 'XEventMetrics') DROP EVENT SESSION XEventMetrics ON SERVER;

Then -OnlySteps '3__XEventSession'.

Collection-time

(dba) Collect-PerfmonData keeps failing

Step output usually includes one of:

  • logman: The data collector set was not found. — run -OnlySteps '11__SetupPerfmonDataCollector' again. Confirm the collector exists: logman query DBA on the host.
  • Access to the path '\\host\C$\SQLMonitor\Perfmon\*.blg' is denied — the SQL Agent proxy account lacks SMB read on C$ of the host. Grant it via a local group or re-use an admin account.
  • No .blg files found — the Perfmon data collector is stopped. Start it: logman start DBA on the host.

(dba) Collect-XEvents inserts no rows

Check .xel files are being created:

powershell Get-ChildItem -Path 'C:\SQLMonitor\XEvents' -Filter '*.xel' | Select-Object Name, LastWriteTime

If files exist but the job inserts zero rows, the XEvent session might be started but not firing. Confirm:

sql SELECT name, startup_state FROM sys.server_event_sessions WHERE name = 'XEventMetrics'; SELECT * FROM sys.dm_xe_sessions WHERE name = 'XEventMetrics'; -- is it running?

Start it: ALTER EVENT SESSION XEventMetrics ON SERVER STATE = START;

(dba) Get-AllServerVolatileInfo errors with Login failed for user 'grafana'

The grafana login SIDs don't match between inventory and the target. Re-run:

powershell -OnlySteps '56__GrafanaLogin' -DropCreatedLogin $true

Dashboarding

Grafana panels say "No data"

  1. Test the data source: Grafana → Connections → Data sources → SQLMonitor → Save & test.
  2. If the test passes, run a panel's SQL directly in SSMS with Grafana's variable values substituted. Most "No data" is a time-filter problem — the panel default $__timeFilter resolves to a window where your collection hasn't run yet.
  3. Confirm collection: SELECT MAX(collection_time) FROM DBA.dbo.all_server_volatile_info — should be within the last 1–2 minutes.

Grafana: "Error evaluating ..."

Usually means a required variable is empty. Open the variable editor (top of the dashboard) and ensure $instance (or the relevant variable) has a value. Some dashboards will not render when a multi-value variable is All — pick one instance.

Dashboard is slow

Open Monitoring - Live - Distributed → Panel title → Inspect → Query. Copy the SQL to SSMS and SET STATISTICS TIME, IO ON. Common causes:

  • Missing non-clustered index on collection_time + instance — dbo.usp_enable_page_compression also creates the expected indexes; re-run step 55.
  • all_server_volatile_info_history hasn't been hourly-rolled-up. Run (dba) Compute-AllServerVolatileInfoHistoryHourly manually.
  • 20+ concurrent dashboard viewers — SQLMonitor's dynamically-parameterized queries help but a dedicated read replica of inventory is the long-term answer.

Alert Engine

Engine starts but fires nothing

Check Alerting/SQLMonitorAlertEngineApp.py logs: likely usp_get_active_alert_by_* returns zero rows. Confirm:

sql SELECT * FROM dbo.alert_history WHERE collection_time > DATEADD(HOUR, -1, GETDATE()); SELECT * FROM dbo.sma_alerts WHERE state = 'active';

If alert_history is empty, the (dba) Capture-AlertMessages job isn't writing there — look at its history.

Where to report a bug

If nothing here helps, open an issue: https://github.com/imajaydwivedi/SQLMonitor/issues. Include:

  • Install-SQLMonitor.ps1 output with -Verbose -ReturnInlineErrorMessage $true.
  • SQL Server version: SELECT @@VERSION.
  • Windows version of the deployer and host.
  • dbatools version: Get-Module dbatools -ListAvailable.