Know your API broke before your customer does.
DevFlow Watch is API observability built for engineers who own the contract between systems — synthetic checks from 14 global edges, response-shape assertions, SLO accounting, and alerts that route through the channels your team already uses.
name: payments-api-charge
url: https://api.example.com/v1/charges
method: POST
frequency: 30s
regions: [us-east-1, eu-west-1, ap-southeast-1]
assertions:
- status_eq: 200
- latency_lt_ms: 800
- body_jsonpath:
path: $.status
eq: ok
alert_channels:
- slack:#payments-oncall
- pagerduty:payments-oncallHit /v1/charges from three continents every 30 seconds. Page on-call only when at least two regions agree something is wrong.
Do not just check the status code — assert the response body shape too. We have seen too many silent contract regressions slip past 2xx checks.
That YAML works as-is. Run devflow monitor apply and the first check fires within 10 seconds.
Three things, well.
DevFlow does not replace APM or your logging stack. It is the layer above — the one that watches what your customers’ code watches: a request, a response, a clock.
Synthetic checks from 14 edges
HTTP and gRPC monitors run from your choice of 14 global edges. Fail-quorum mode means a single edge dropping the connection does not page your on-call.
Response-shape assertions, not just 200
JSON-Schema, JSONPath, regex, and a Zod-style shape DSL. Catch contract regressions before your customers do — even when the status code is fine.
SLOs that mean something
Error budgets, multi-window burn-rate alerts, public + private status pages. Page on what your users feel, not on every flap.
“We caught a 700ms regression three minutes after deploy. The previous tool would not have noticed for an hour.”
Try it on one endpoint.
Three monitors are free forever. Five minutes to first alert. The quickstart walks you through it.