Monitoring and alerting for SQL Server, Azure and everything around them

Every replica.
Every wait.
Every pipeline run.

Excubius reads from the monitoring you already pay for, fills the gaps itself, and tells you before something breaks. It never touches a server; it watches, and it warns.

Praemonitus, praemunitus — forewarned is forearmed.
ag-prod-01 · Payments · Production
42 healthy3 warning1 critical
sql-prod-01
Primary replica
Warning
CPU71%
Waits · PAGEIOLATCH340 ms
Log send queue12 KB
Linked server WAREHOUSE_DB2 timed outAck
sql-prod-02
Secondary, synchronized
Healthy
CPU18%
Redo queue0 KB
Last backup42 min
No open alerts · last check 20 s ago
Listener sql-prod-lis not resolving from the app subnet · 4 minSeen by 2 · ticket OPS-214
Reads from what you already pay for
New RelicAzure MonitorNinjaOneWazuhDefender for CloudAWS CloudWatch · next+ the Excubius collector for what they miss
Products

One suite. Named by what it watches.

Start free, add the products your estate needs. Everything shares one entity tree, one alert inbox and one status page.

Excubius for Databases
SQL Server with Always On awareness, Azure SQL, Managed Instance, PostgreSQL, MariaDB. Waits, queries, linked servers, SSRS health.
Built by a DBA, for DBAs →
Excubius for Infrastructure
Windows, Linux and macOS hosts over WinRM and SSH. Azure VMs, network, platform health and Data Factory runs. AWS next.
Excubius for Security
Posture score per area, findings and events from Wazuh and Defender for Cloud, each switchable, recommendations optional.
Excubius Advisories
New CUs and hotfixes, OS and app versions, end of life, certificate and secret expiry at 30, 14 and 7 days. It advises; you decide.
Excubius Synthetics
Checks that run from inside your network: directory sign-in, app to SQL, portal render, listener DNS, linked servers.
Excubius Alerts and Status
One inbox with suggested actions, seen-by, assignment and one-click Jira. Slack, push and email. A public status page per entity.
Coming next
On the roadmap, in this order
Public roadmap →
AWS
EC2 and RDS hosts, Aurora clusters, CloudWatch as a source, alongside Azure in the same entity tree.
PostgreSQL clusters
Streaming replication and Patroni: primary and standby roles, replication lag, failover and switchover history.
MariaDB clusters
Galera and replication topologies: node state, flow control, lag per node, split-brain detection.
Principles

The guard who raises the alarm, not the one who fights.

Warn, never act

Excubius is read-only by design. No agents that patch, no scripts that restart. Your change control stays yours, and so does your audit log.

Cluster-aware from day one

An availability group is one thing with two roles. Primary and secondary get their own metrics, thresholds and alerts, and a failover does not look like an outage.

Baselines you can defend

Every threshold starts from vendor guidance, inherits down your entity tree and can be overridden per resource. A catalog of 101 checks in 13 modules ships in the box.

Alerts

From alert to Jira ticket in one click.

Each alert carries the probable cause, a suggested action and the state the resource was in when it fired. Mark it seen, assign it, or open a ticket already filled in with everything Excubius knew.

  • Slack, push and email, with a dead-man heartbeat so silence is itself an alert
  • Seen-by and assignment visible to the whole team
  • Jira status flows back: Done closes the alert
InboxUnseen by me · 3
Listener sql-prod-lis not resolving from app subnetag-prod-014 min
Linked server WAREHOUSE_DB2 timed out after 30 ssql-prod-0122 min
Pipeline pl_nightly_load ran 38 min over baselinedf-prod-013 h
Probable cause
DB2 host ibmi-prod stopped answering on port 446 at 15:02; the AG failed over to node 1 at 14:58 and the linked-server login maps to a different credential there.
Suggested action
Verify the linked-login mapping on sql-prod-01 matches node 2, then test with the catalog query the check uses.
Pricing

Free up to 10 hosts. Then per host, nothing per metric.

Compare every plan →
Free
$0up to 10 hosts, forever
The whole product for labs, home servers and small estates.
Core
$8per host / month
Infrastructure, Alerts and Status for one team.
Pro
$24per host / month
Everything in Core plus what a database estate needs.
Enterprise
Customannual, from 100 hosts
Everything in Pro for MSPs and regulated teams.

Ten hosts free, forever. Thirty days of everything, no card.

Self-hosted in your own cloud subscription with Terraform, or on a Linux box you already have. Point it at a source and watch the inbox fill.

Get your free key