Slack
Incoming webhook per channel. Schema digest, schedule run failure, agent offline. Click through to the workspace.
Slack setup
Optional · Alerts
Data schema alert channels for schema changes worth waking up for.
A data schema alert is an opt-in layer on top of the documentation stream. Route Slack, email or webhook notifications for the events that matter: a likely breaking schema change, a failed scheduled export, an agent that went offline.
Transactional email to a list of recipients. Same digest content, formatted for inbox preview and accessibility.
Email formatDelivery log
Every alert attempt is tracked with channel, status, retry count, last error, and next attempt time.
Schedule deliveryRules
Match the event. Hit the right channel.
A rule pairs a trigger with a destination, where the trigger is schema severity, the status of a scheduled run, or the health of an agent. Destinations are channels, and nothing stops several rules firing on the same event.
Schema severity
Info, Attention, or Likely breaking. Pick the threshold that wakes the on-call.
Object-type filter
Include or exclude tables, columns, indexes, views, routines, foreign keys.
Schedule failures
A scheduled export failed three times in a row. Surface it.
Agent offline
The agent missed the scheduled scan or has been disconnected for too long.
Why this signal deserves to fire at all
Structural change is the signal query performance dashboards miss.
The reason a schema change alert earns its place next to CPU and latency dashboards is that the runtime observability tools never see the change itself, only its consequences half an hour later. The two pieces below cover the reasoning: the first on how the gap between the migration folder and the running database opens in the first place, the second on why watching that gap is a first-class observability job.
Schema drift detection
How silent structural changes make it into production without a migration, and what a diff engine needs to detect them cleanly (renames included) instead of drowning the team in false positives.
Read the pieceDatabase observability, but for structure
Query performance monitoring answers "how is the database behaving right now"; it does not answer "what changed since yesterday". Why structural change deserves an observability layer of its own, and how it complements the runtime one already in place.
Read the pieceReliability
Retry, audit, prove.
Every dispatch is a row. If the channel is down, the cloud retries with exponential backoff. The delivery log shows status, attempts, last error, and the next attempt time.
Exponential backoff
Up to six attempts with growing delays.
Per-attempt audit
Timestamp, channel, status, attempt counter, last error.
Same event shape
Retries keep the same event details so Slack and email recipients see a predictable alert every time.
FAQ
Data Schema Alert: Slack, Email, Webhook for Schema Changes
Which databases are supported? +
PostgreSQL, MySQL, SQL Server, and Redshift. Any of them can be alerted on schema changes, schedule failures, or agent health.
What events trigger a schema change alert? +
Added, removed, or modified tables, columns, indexes, triggers, views, and routines. Each change is classified by severity: Info, Attention, or Likely breaking.
Does the cloud read my database rows? +
No. The on-premise agent runs the schema scan and sends only structural metadata to the cloud. Row data never leaves your network.
Where is my data stored? +
The control plane runs on dedicated virtual machines inside the European Union. A DPA is available on request.
Can I send alerts to Slack and email at the same time? +
Yes. Create two rules with the same trigger and different destinations. Generic webhooks are also supported.
Wire it up
Pick a channel. Page only on the events that matter.
Slack or email, with a severity threshold and filters on each rule, retried on failure and recorded either way.