Data Stays Local. Alerting Lives in One Place.
Monitors queries your metrics, logs, and databases right inside your network, brings alert rules scattered across environments into one place, and hands alerts straight to on-call and AI SRE.
Per-service 5xx requests per second
sum by (service) (rate(http_requests_total{code=~"5.."}[5m]))
One rule, every environment
Edge engines run in your data center, cloud VPC, or Kubernetes and query data sources locally, so raw data never has to leave your network. Bind one rule to many data sources by name pattern — change a threshold once and every environment follows.
Flashduty Monitors
One rule
API 5xx rate
datasource: prom-*
Plant A
East DC
Overseas VPC
PrometheusElasticsearchLokiVictoriaLogsClickHouseMySQLPostgreSQLOracleSLSCLS
Data stays home
Engines query locally, no copying raw data to the cloud
Engine-loss alerts
Get notified when a whole cluster goes silent
Automatic failover
Engines in a cluster back each other up, no duplicate alerts
Alert on thresholds, on data, or on silence
Threshold, data-exists, and no-data modes, with Critical, Warning, and Info evaluated together. Recovery can be confirmed by its own query, not just a number dropping. Stuck on the query? Let AI write it.
Edit alert rule
Query A
Write it with AI
Per-service 5xx requests per second
Import rules
- Prometheus Rules YAML
- Nightingale v6–v9 rule JSON
- Flashduty rules YAML / JSON
Imported rules start disabled until you enable them
Every kind of signal
Over a threshold, showing up, or going missing
Recovery you can trust
Confirm it with a separate query
Context in every alert
Template variables carry values and labels
However many rules, see what's firing at a glance
Rules live in a folder tree that rolls up active alerts and failed runs on every node. The entity tree discovers hosts, pods, and database instances from Prometheus query results; attach rules to a group and new instances pick them up automatically.
Alert rules
Entity tree
Status at a glance
See which group is firing or failing right on the tree
Rules follow entities
New instances get group rules automatically
Permissions by folder
Each team owns its own slice of rules
One place to query while you troubleshoot
No more hunting for each environment's address and credentials. Query Prometheus, Loki, VictoriaLogs, SLS, CLS, and SQL databases from the query workbench — or click “Query” on an alert to open it with the query and time window filled in.
prom-plant-aRun
One entry point
Every data source in the same view
Arrive with context
Query and time window prefilled from the alert
Easy to share
Query history and share links instead of screenshots
Alerts reach on-call, AI SRE digs in
Monitors alerts flow natively into Flashduty On-call for grouping, routing, and escalation in your team space. Diagnostic sources like Redis, MongoDB, and Kafka are available to AI SRE while it investigates.
Rule fires
Edge engine evaluates on schedule
On-call routes
Grouped, then sent to whoever is on duty
AI SRE investigates
Queries metrics, logs, and diagnostic sources
Auto-recovers
Closes when the recovery condition holds
RedisMongoDB
Kafka
Native hand-off
No webhooks; alerts arrive with labels intact
Noise handled in On-call
Grouping, silences, and inhibition in one place
Repeat notifications under control
Set the interval and a cap
FAQ
Less noise. Faster recovery.
Start free—no credit card. Book a demo to see your alerting pipeline end to end.