Product
The platform that watches your vendors
RELIASTRA monitors the external APIs your product depends on, correlates their failures with your incidents, and produces independent evidence of what happened.
Who this is for
SaaS teams whose product breaks when a payment, auth, messaging, AI or cloud API degrades - and agencies that operate customer infrastructure and need to show clients exactly which dependency failed. If your incident reviews ever end with “we think it was the vendor,” this is the record that replaces “we think.”
What problem it solves
Your own monitoring reports that your service is failing. It cannot say whether the cause is your code or a provider three hops away whose status page still reads “operational.” That ambiguity decides where engineers look first, what gets rolled back, and whether an SLA claim succeeds.
How it works
1. Monitor every dependency directly
Add any HTTP endpoint. One scheduler dispatches one check task per dependency per region through a message broker to workers - never inside the API process, so a busy dashboard cannot delay a probe. Every result carries its region, timestamp, latency, status code and outcome. See dependency monitoring and monitoring docs.
2. Declare incidents by quorum, not by single failure
A single failed request is never an incident. Declaration requires failures across more than one region inside a short window; recovery requires consecutive successes. The same evidence produces the same verdict everywhere.
3. Attribute with a deterministic engine
Your incident history and the dependency’s observation history share one timeline. The correlation engine reports overlap with confidence levels - incident evidence, never claimed causation.
4. Generate evidence you can submit
Timestamped, checksummed fault reports bind the observation window to your organization. A public verification reference confirms existence and integrity without disclosing endpoints or credentials. See SLA evidence and evidence docs.
What data supports it
Live aggregated posture for public vendors on Track, the published measurement methodology, and the research agenda that states what we will and will not publish.
What to do next
Start free (3 dependencies, 1-minute checks), read the quickstart, or compare plans.
Know when your dependencies fail. Prove what happened.
Monitor the external APIs your product relies on, correlate their failures with your incidents, and generate verifiable evidence when a vendor causes downtime.