Overview¶

The SQL Server AlwaysOn probe monitors the health of SQL Server AlwaysOn Availability Groups through the SQL Server Dynamic Management Views (DMVs). It reports per-replica role, synchronization health and connectivity, and per-database lag, queue sizes and throughput rates. One probe instance monitors one SQL Server instance; add more instances for additional servers.

Collected data:

Server reachability
Per-replica role (primary / secondary), synchronization health and connectivity
Per-database secondary lag
Per-database log-send and redo queue sizes
Per-database log-send and redo throughput rates

All metrics are emitted under the senhub.mssql_ha.* namespace. Replica series carry an availability-group and replica attribute; database series carry an availability-group and database attribute. These attributes also act as filters in the Web UI Sensor Builder.

The probe also emits a db entity for the monitored SQL Server instance so that the topology view can build a node from it — surfaced even when the server is unreachable or has no Availability Group configured.

Quick Start¶

Basic Configuration¶

# probes.d/20-mssql_ha.yaml — each file under probes.d/ is a YAML array of probes
- name: mssql-ha-prod
  type: mssql_ha
  params:
    host: sqlnode1.example.com
    port: 1433
    username: monitoring
    password: "${secret:mssql-ha-prod.password}"   # OS secret store; inline plaintext is auto-sealed on install
    encrypt: "true"
    trust_server_cert: false
    interval: 60

The ${secret:...} reference resolves the password from the OS-native secret store (see Configuration).

Multiple Instances¶

Monitor several SQL Server instances with separate probe instances:

# probes.d/20-mssql_ha.yaml
- name: mssql-ha-dc1
  type: mssql_ha
  params:
    host: sqlnode1.example.com
    username: monitoring
    password: "${secret:mssql-ha-dc1.password}"

- name: mssql-ha-dc2
  type: mssql_ha
  params:
    host: sqlnode2.example.com
    username: monitoring
    password: "${secret:mssql-ha-dc2.password}"
    encrypt: "disable"   # test/lab instance with no TLS

Configuration Parameters¶

Parameter	Type	Required	Default	Description
`host`	string	Yes	-	SQL Server hostname or IP. For a named instance use `host\InstanceName`
`port`	integer	No	`1433`	SQL Server TCP port
`username`	string	Yes	-	SQL Server login with read access to the AlwaysOn DMVs
`password`	string	No	-	Password for the SQL login — reference a stored secret via `${secret:<name>.password}`, `${env:VAR}` or `${file:/path}`. Inline plaintext is auto-sealed into the OS secret store on install.
`encrypt`	string	No	`true`	TLS negotiation: `true`, `false` or `disable` (go-mssqldb convention)
`trust_server_cert`	boolean	No	`false`	Trust the server TLS certificate without validation (set `true` for self-signed server certificates)
`interval`	integer	No	`60`	Collection interval in seconds
`timeout`	integer	No	`30`	Per-cycle connection and query timeout in seconds

Metrics Collected¶

Overview¶

Metric	Unit	Description
`senhub.mssql_ha.up`	`1`	`1` when the last ping reached the server this cycle, else `0`

Replicas¶

Each replica series carries an ag.name (availability group) and replica.name attribute.

Metric	Unit	Description
`senhub.mssql_ha.replica.role`	`1`	Replica role (Primary=1, Secondary=0)
`senhub.mssql_ha.replica.health`	`1`	Synchronization health (Healthy=1, otherwise 0)
`senhub.mssql_ha.replica.connected`	`1`	Connectivity state (Connected=1, Disconnected=0)

Databases¶

Each database series carries an ag.name (availability group) and database.name attribute.

Metric	Unit	Description
`senhub.mssql_ha.database.lag`	`s`	Estimated secondary-replica lag behind the primary
`senhub.mssql_ha.log_send_queue`	`By`	Log records on the primary not yet sent to the secondary
`senhub.mssql_ha.redo_queue`	`By`	Log records received by the secondary not yet redone
`senhub.mssql_ha.log_send_rate`	`By/s`	Rate at which log is sent from the primary to the secondary
`senhub.mssql_ha.redo_rate`	`By/s`	Rate at which received log is redone on the secondary

Filtering (Web UI Sensor Builder)¶

The PRTG/Web UI Sensor Builder exposes filters for this probe:

Metric Type (category) — overview, replicas, databases
Availability Group — pick a specific AG
Replica — pick a specific replica
Database — pick a specific availability database

Requirements¶

SQL Server with AlwaysOn Availability Groups enabled and reachable from the agent host (default TCP port 1433).
A SQL Server login with permission to read the AlwaysOn DMVs — VIEW SERVER STATE is sufficient. A minimal-privilege account is recommended over sa.
The probe reads sys.dm_hadr_availability_replica_states, sys.availability_replicas, sys.availability_groups and sys.dm_hadr_database_replica_states.

AlwaysOn required

This probe reports Availability Group replication health only. For general SQL Server health and throughput (batch/transaction rates, buffer cache, per-database I/O), use the SQL Server probe — the two are complementary and can run side by side against the same instance.

Outputs¶

SQL Server AlwaysOn metrics are available through every configured output — OTLP, Prometheus, and the pull formats (PRTG, Nagios, Web UI). For PRTG and Nagios, query the probe by its configured name:

curl "http://localhost:8080/api/{agentkey}/prtg/metrics/mssql-ha-prod"
curl "http://localhost:8080/api/{agentkey}/nagios/metrics/mssql-ha-prod"